loffeld-et-al/IMG00002.GIF100644 23423 23424 1656 6220461260 13504 0ustar luziguestGIF89a-@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,- H*\ȰÇ#Jl@'b"ƎAfTQő ;oeΊ@wx0gF>dȳ7> jBBc4ǫL%~5ҫS3L;UWnjKݻx;loffeld-et-al/IMG00003.GIF100644 23423 23424 2557 6220461260 13506 0ustar luziguestGIF89a(@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,( H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˗ , fB5gҴɳρ9I0hLH RO76UaUI^dΩ5~`٢=+5ڈlJ6+Ui5~6c_o5cLm6d`73~(8V٪NŘyQ,uܡ6lʕvl=u[g{?W]?gfvsOXuϴܴzRm=iG.|qI\v%X}dvx%HUUytavf(Zg!:nȞ]~'~'eiݹZ_m){!B^A_׃6(7]Fa ZUi۩x롷 ZnX1XWd+qM#|9c bwg|r9=jWMfї;N饔]ME~v_NZWNYhP*H"u6+L%(8k|Ȗ'o6+mM"lDrk覫!;loffeld-et-al/IMG00004.GIF100644 23423 23424 7042 6220462174 13506 0ustar luziguestGIF89a@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!, H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˗0cʜI͛8sɳϟ@ hѡH*gҧPJΩV.ŪW\:َaϪv۷ *u 7"ݐmE{Tެ'M:x/;qU ! 劊+L\j~ڮң66`뛯>|:5egf*X^Ժ}q;ŏȪ( Qe;{|(Ol02v/,v^oTBsB&bx(fъ'>n}rvesx<{8<6ŵ,&x:ʛO~w_~{]2v^,7y==/+~^<իO$O%_:I~ 咦<\%ڇ$i2nKnKSALᛉ9u=qpPK޻r<)鄬{aÔFApP$kF6#B]8.?, !!$0Sdoȶ(v劭Ũ6,2胢kd(G̉aHő$AhLX?E:<_$'IH̤X.Nz*(GIRL*WV򕰌,gIZ$!sK,Q0;'`vD1q͌f70O~.wB/N,\*?VosC%9s0u\gN\#=scM} AOܤOFLp7Tfh2Pg6PlԛЀJ>atЌ~ 5 DӉȻ]KJ={3 ۣ9MoMwA3|')J]X q4VSR+jhZhatjBᙺ Ԣ>'1; IT^nSYDcSbК"ӫe,WǨ<|}]V~7u4>i$;aEXw,(p/aּZqJ,k!:Wfؽ:dpUEhm̻E}'Xacr\500]Q''/]kLZ,rY␊?o8g-"dJye-=Mr$74CVj,F tlˎڙH l Эf)8[e*:z:g:&ne1L1% Mbd#jhKMmM}'E`SWY+NGM*t5_|VuJa^`bNtB'*nL h9kg6SVbܖY>s?I׊sU߾/z_|/aETnsH'8҂Z޲Y5~zMf&w!m^QSU}إq?ge4}]pww%H yM[tq'Ԇ:wZ!Rq]HXE`"x^n'a^UZx#vu~foB(@7ܐD(ё QSވcθߵ8Xfwz tfbyngdbǽF$Z5`y݃b\<% %]%iiX&w~ ړd(fHdU* :jXVwdJ^,vl?66**5Yq~fR扚]vrin[A켍f;B,%9h>dXpK 81ŋ;_z]_q`Fn, 2,k%ȨÍl-j|pOb6O1n0>L(?v;.cm0\qaoAnU'e6>+Lv]o\Nx/Uv=Vԙɺzٰ]yCL(kg.yem擓n:Y[yFzYN;RYyW̓n{GOWogw/o觯/mϏ\=nobR?DS m.g9{ಯ p5QGw6cܜR28*ֲ 8ml 4\xp*4l 9Pg̱rF""2D RA#BٻbA}S'dM@4Z6H%qlNF)tj̔9qطQcC9 ժ8 Xlx0HkdSÂ(;:'L(=yImN0{t"$CI⋊#$N"Q fH&2a}ր<%D2e:dr)rI[_ P6J5k7znFn'\ @[riVs [ C #hF4$miПtIE=RԨN;loffeld-et-al/IMG00007.GIF100644 23423 23424 1665 6220462174 13516 0ustar luziguestGIF89a,@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,, H*\ȰÇ#JH1"F96 )H$,pe”.]rJ,fG*{idL< 4gQH}B"I4RX-NZSƕJ zD[JYb׺e˷߃;loffeld-et-al/IMG00008.GIF100644 23423 23424 2226 6220462174 13511 0ustar luziguestGIF89a@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!, H*\ȰÇ#JHŋ3jȱǏ C9IH\aJ/[ʜI͛8sɳO&.|)t`LDOVydңM]UشJZІ^ `ʰ neՠm;up.ǷqKZZ LVja{xIb3> 3_ɉBLC_ UԶn7מ} lf]6ltsj=_ݙ5ଇ+Yc^tO?߾\At,l{om2GO|is<=_Z 8t|⩷o-I 9leއf]"~a#ى蓋?cx3h8IGlx3+S?Q [o!yd7jə`|I7W%W[HW|HkeIq ֔:ךg>xp덗&"Hh'#r8v6|$sh#8WeiA8 OB呰v42dYyoZfprg|e>&+VT" gJkyU2'6{>B^_V(ꥁHfeyAZꬤܪ{96+z"o_i'02{#)=nl>Xi痶N~j)dZ߶V׫!hhjV]n5 /**-F/톱v(p·z?nʺ~Xܪ!*HJodK_ZbUPRuv~%-gW}iFT W aAHg^Uah|#~M\]ZFB?ⶤRyXs6&ؠBUq]ULfcif)=(ޝc_j١^x|d-צ) jo\9'壅8蜛Ng%w!" 2x`NiCΪ'fr $ި+*稭"n '6f^~"yG66 V.Ȧ{dJKmeHYo]u~)jﭶ|R"dI"c+v3_ܢ͔QX{ ÈF%XvKxEG>U(~*j‘̂H(ᯎ_5jdX,ؼ?"I:rڤGůos{d#KrC讐HKF wj\R|~FU ѓ=xDE.e +9hN6pL:xγ>ϖy2 A#І>dF;;3s(Ip̖!3 Msڃ]( 9X*K +G,ʷf٩!lar:ΩR;וZ &u"_sTKzEW{>r54֭Qu}5WFۣ`ї55[l@rb{G:WQMo[1(&}KW#vV`άRxh)L~-;SebE}wUZ-waqiqs K77:L#/v;|\pr_e>^ D^G7bcwfحwkwK=ǸSoM<-ʅ E3񐏼'OyH;loffeld-et-al/IMG00011.GIF100644 23423 23424 7147 6220462176 13514 0ustar luziguestGIF89a@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!, H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˗0cʜI͛8sO< =УHӍLJJaԪ b5խ`"* ٲ.T۴o9gw d^sUWEҨ-F>(`lxrVf-vL牧vxP·Y+zVsN};qk;1ջ7.ʗ*6zm嶧>lYjQ&w}G^[f~aԁfWt ^htc—bxBw+fg"%[`lKIdmcx@¸eI?V!1F) 6#-X >Ğ{"gTY喢M!C$mIF禘rJ&eyeX]%+ffk.izjP(d梘MJAgYh*}gflr褌FBc_Tb:z^)_> %:%V_W"}gYɶ**+V*5~dlm &v:yQ^rzS蟧J豲YnPܢۯyKo,uK|*N'SwS pk`b qoRq%ktap(oM8r˝LܫDQJi:ȳTI/.=h]on2N}R?5XU2hϘua~l[wI&77G.gΝMw砇.zeLL7{xVUi|~qٺ0Voo0]=7Os~ɲ}S]&h 2ډ@=ʛVT2Gꨥwr| %?~ޠĵk[J&Ăp묶:m!%H]W=v-ä́G0 !dEyS`83m2<0=s݋XD:uSgHp- IEeJ_GMnT[L W|a7>ણv"iMɮ"l"Y.Ԑe3Pu ) m%JHWJ2aFl%K&S̔Lw7=3%OJԢHMRԦ:PTJժZXͪVծz` XԱUf%#>fPhM'h7MUw~}5r鰓WX{lc:%vc5\+pV5j/'Xd5ʠ5%LI2+nSk–V{8Rk(!m-Eμ*U_oI8eEn+:w,.]?:^<EvR= {Zn锁g$hHílTnB]{ZX>$dK?ރS:aʷd :,vTkK^ҧ#SF7o8 ؞̞#}+huwb [bPxZ$qA\(P5ǐ.fk-\$s]"g 3̡ᱎ{3p;jG3tY/ךN[sf:3Ze-ωMI7.{AXФ|{7`:1ni\ if[QX dmRD 6SK֬LvEyf7ubV@Gee$e&u;òqVWǾZYf_؋lc.m|{yO([u4Qyޜ)+\^Sʾ/\q`W5V(uOUewC% 9G϶{z};loffeld-et-al/IMG00013.GIF100644 23423 23424 4711 6220462200 13474 0ustar luziguestGIF89ao@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,o H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˗0cʜI͛8sɳϟ  ѣ7E#PJEtՅUja֮[KزQϢ]֧ڶFKܺsyI \,5c'xYKx_* ={{CX{ ]wGw7w9X[8Q!vBȟuN衂'^aojTF+ h\$'c3#ѥc|W8Nh#~$"`-FގォYaZV^|q,,9;%_/ܗ^>+i:w5`߼˟ frsw'M]˺i×oߘۼzE,*_+' MgTJŞ/cӒ쏀r_;Ѝcb Q&=Cx.p# %@>(3Zd#b-Q˒s$趂)#=C0qb4G9QZ2xEKF3eD&b^ ׎ir0<uּ#?;3C <*&NA[˛29O %wo_]gl <%b0f7nj,ʝ 7 A/}*덙z7L(NWL;loffeld-et-al/IMG00014.GIF100644 23423 23424 1661 6220462200 13476 0ustar luziguestGIF89a*@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,* H*\ȰÇ#J@!b1Ǐ C9ɅO쨐%K/_Rl(`L5a2if΋:Ul"΂)etfΧL4ЦHgZ J\4C%~\9lBAs^H-kՑO% 5^ˀ7g,vtѽ=jg.#<6tç:~T}6=|<Dž#{m%i uV`WFu~镗~`cقZۃn8]&"~%b;b-/jg#Ysxz Fal*^dnƱFevI:ԚNUۊD^ӎ%!ff2cpWn y9#q g"fM gn)^NR z裐:餔ffhs]j&XjJS؉z]vZԨVhAYdo4b$ZDd왞#Ni޲{+f-jev+许V.5{m]gYӯeەɟ]m(z>^Ċ GFy+n+Hޑkb(|5ljɣr*,If+̜F;!i4豑,N 4l4GK̆L'gOϋ#2S5T5{6[.6/icYߥ]0*df+o뮍 _S09.2K oιI}wֿ"{V{ "\.3YS^˴v(WQGCIxaіx̜#ChYΎ<׵7:h_B[XA>V7.9VQ#DT%lWo<⮎d SxAn1H)~%SOFAQ䋦<(2-AKYAyŇ1yLf:3+|4;8j2*̦6nz g6;loffeld-et-al/IMG00016.GIF100644 23423 23424 1646 6220462200 13503 0ustar luziguestGIF89a*@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,* H*\ȰÇ#JH ƅ92 )ȇ$,p%”.]ҢJ,-nr&G=3ACs4jhML}:m)uIONrʤS!f Q`;D[۷pJ ;loffeld-et-al/IMG00017.GIF100644 23423 23424 5755 6220462200 13511 0ustar luziguestGIF89a@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!, H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˗0cʜI͛8sɳ'J| UtѣH}4ХMJJ5%TE޼Uׯ`!6 5,̱Lͪ]k6*Y/ݦKQPK@{ LÈ+^̸ǐ#*#ʘ3"Πg mq4ӋMz_3-lenyk.wp9·{} ebS3︵ϭ=6瓧w(^8b8O݇ۺ 1 qD}TqZx!X|yYq:"m%vsb~w"rgo1҇\/bYM6-"%u$#ֈ#z QnhabVKcYz 1 ɟf7T>|6%l٥Vc)')g|4g)`U=ez9ZjۅJfeڤz(rcu)Q\ݩͺ(iV^{D+SWl qkl,N((훜TkW壓"kr6i*s`jYBjfh+NӿZܢ.tR7RZ'鯳~ڦBZZ\("kX+[4&»~ҩql}̯FL4n`"o:|:~fp_էbggS܇nѿ-c >i\7_Q=rݟ04#L:'H Z̠7z GH(L W0 gH8̡w@ H"HL&:PH*ZX̢.z` cC29$ wFOllci$Cx̣Gϕq|cGS |cGErz#SF9ui[$I$1g3AIH',ItuҒ 2Qr~:nSϖ ti4R\ܥ9aJt7kʛ FMO!'*Y3H$ɯHAGŊr(I,gI w#Wsh*2P9dK&ʬM4fO}5&&y4RI'0O9.PHm<>%Kh38vsgO)iftlhJwUJu9&PD XS4JIqXj=<}yꕡ[ Bzv/[Xt1llj> VBʼnGן>Y]F cy.EXW.h=mdV+NUWm{[ Y9L߼dӜZrsY7nǭ.aUl,rZfQ EBop;n~ /g'W65N١T& X1-s^= 0M[ 0-` c4v]:r `%}wVFAH~wG&r3 rcJIorJ9gqv< s!L>>'Q&/TsLB;πMBЈNF;ѐI0͖!3C9swt \gQԨ>tSUڄ~5c-k ҺdzK(Vx]&Ye\~,8Wj}z]f%2eڢ5NU.pCږyip>5DZګޞbIM*ޕdu6%eTS0M3#+gnx XYtXk2}[jjW>ld^qz<]b=لg7u|{Ԃ3Dܪ=YwAJUKo_npӕu>67w=hZ sI]Jw|ޭr%Wj'^)GOқ|37WcOc@;loffeld-et-al/IMG00018.GIF100644 23423 23424 7375 6220462202 13514 0ustar luziguestGIF89a@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!, H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˗0cʜI͛8sɳϟ@  QDJv!|MSe^yYf'j%՚gT  .}i}w>Zm.d[R꧖:t2J}I)@:*z&av;Jj(&L (yI"Goo>ɤwFw妢rUښf[lxҞJVik(bo9.mu;K|zء:rx:~i˧9*11jol^;.Vk/3`j/FXq\mޞSbg05# 4M,ڳHR[.zXOWoKkfs[($u=j-dsy&lkR%!mdH*zrֈtb v LxǤ)rn"P;\oGt.̞ ת#&Կ|I.W<>bow/oQ<>'2`_]@L:'H Z̠7z GI Wqɏ0bHð8M@ _!%/<e%:q4M|w6*29*շXe*lQI]`8vvKzgl߸ ,rkL#D gȺ9ghr*D$Yܓ瘅͑W5'BQO`:w&Wrt2HUaR8HXX8$$-.4'F%36jd{̤">kx<JaܢRƜ0X5_ M8;HGEKFkIY<:S%hHj17ꐝ%&2͢c-nSO*mFeHK84`d^b(4.:ŋl2ÛT|QӔ.PTJժZ+ɺ@ͩUDroXJֲhMZֶp\J׺xͫ^׾ `KMb:d'KZͬf7+r|3l*ZÈz(25rc1A<-W£ˎ!rmʌ"U-3[ݦͦؖO"ojmŝ&quC-EZ"RKJNwBs|'TA6e,eV/ju]?ݞ*.d'*;@fwXh_0D)t[txGaOuyT<^V8+0IOsZ`D cn2܂oGeyiפU0Pq7jDvczf;W*K,Cy=ڱS$TLQSS#,`w.)eCRYq$%Hwv)}H**#Puq-}BlQՎeTJ {M7^\'XZY뭌믅˶u.V#9EP_ܛr'%\xktb&{ۗ{&u.-%V#Y7Ux;˳5|;Tk}ld?O#'ٺu}yi#/6/ sr7p> Ѓӡ= gųN 5@v_~ofoLJk;=w=|r;Q`._JrOS3<8#Y~nON; '5jI~$/w;1HE_̀򉧩r9`׻x/|}D~ ;ЏO[Ͼ{OOOoG=1&TPVlVQBX]x8onzQ/|XHDH;N^X!hH#9*h n.0Hk2L7d-x`.l;%ƂEI("m<#dф.7W_ xnn gf7a惷f\qV(2U0hin5Z=Wc`ui\ޤ_Mq (8Ed&aFTJd_C`ʶma$䈵a[bز\v953(V/ttk2unhd*ﶊjHc*Tk&g& .tmƂ6mc8y:a8 r֌f#nc2H&mSI٢^hŌ掩6hdxLƎň`8d4iɘUzv‹=T0[x&89fOgsiZRNxMlji-I5Ј5l6G)imsRVm[72dn#nDPO>5r%oQb<9pƒodYGf9+Qpx%`WqriZpIvcni@gi|wx32?Ay7X-n";t敷xr yEqI$)u`t{P鍬h/u)b( q4I"ٛSz#YyU;loffeld-et-al/IMG00019.GIF100644 23423 23424 2773 6220462202 13512 0ustar luziguestGIF89aT"@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,T" H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˔b0&8Ex3ɞ>J"PEu& tӧ4Be)uի#b?C" ؄=wÛj)*iU ޶fe EX_%xm ׼YndÝ- ǕM=QpUwL͚/²m̐6YФscMw!<9Gߺ ȥ-߲ͭݍf?>|mx'ce$gAR~DZ PǛ{顓Bhu$HxYyeg6hzXY|ngҩ/M`aZ9:ex6&>)Ra6Xri& ۦf.穴.7dU nkh\,)ng06\+l/,Ɠ:l䰮2B ý|,k,GcqKs;rʚol4jK)RkÕ03=-)g5`M]|56w.m7gtpZo=o-eICطqvsdVxOp 8aV`S}9Y}`G%>u-7M5qb?ns{K{+dSvak`Xϖ'U#uy3_Mc^H.,=n`( Z̠7z GH(L W0 21gj8!OphM8Y=$lK)I_.j=p߸H bGKaH@%`$"X=*tc/QK ($1 cjx񑉌Z`I򒠃&}ELz2 :Rguۊf(hHS2' ,7iW*‰ei[2.REbO2!%ЌfŖIM0'u6O59ژ kFl /I^ֲ"1il':>roCm+e h71 L#D>:V[VJщZ1fF EZt %=SАBQJzҖ&1.ڜ%Ӛ5E Ӝ@ PJԢ5";loffeld-et-al/IMG00026.GIF100644 23423 23424 1645 6220466524 13520 0ustar luziguestGIF89a$@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,$ H*\ȰÇ(@D+fȱǏ78Qȅ'C<Ŕ,]bsL4m,Iʊ=uR pɝ7]rN9m3OK_Y+ɭNB"٫ϞX۷p/;loffeld-et-al/IMG00027.GIF100644 23423 23424 1556 6220466524 13522 0ustar luziguestGIF89a@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,K H*\ȰÃ"F|`E ^(pbƉn7RaHG4iʁO^|ǎgr$Isϟ;loffeld-et-al/IMG00028.GIF100644 23423 23424 1657 6220466524 13525 0ustar luziguestGIF89a-@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,- H*\ȰÇ#JH?1^Q`DžcHO^D2!I-|x$G1;Lscσ:ET%J7a"4$ѥ?)(ǫWK"խZ3 UתR)&}ZVY^wu9ݻx ;loffeld-et-al/IMG00029.GIF100644 23423 23424 1556 6220466524 13524 0ustar luziguestGIF89a@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,K H*\ȰÃ"F|`E ^(pbƉn7RaHG4iʁO^|ǎgr$Isϟ;loffeld-et-al/IMG00030.GIF100644 23423 23424 5673 6220466524 13520 0ustar luziguestGIF89ab~@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,b~ H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˗0cʜI͛8sɳϟ@ JѣH*]ʴiN<*`ՂfMxթׯ`ÊkMZxu@cʝKW(ܙk½߿B ^̸cZ/m_[J~̹g'<reš?^͚hj1۸}ڶs N\]C&?̖۷Vn.NyW]1GNuwx_uETi߂ WV_}I{"a vj X7YQ~⋏Ezɵ{эb٧dH"UƜIF)KXfyq`#aihlp)tix|矀*蠄j衈&꧌8oVZih,~v$a%^M(JCF*R"*,MYA`6XFRlfj Dm24q ҹi_nFqبbk~zYSWnZھ޺pM+j6Eh nIZ5jn<n0q % pi#-2RȂb>%r0k+EMS\ YB}lg 18 h: 2EG-tXF2>b ʎ+b y; .cx?u0E #;2H m$ A `DN6redbSW #-yYI{dMG'AVqe$GS$i/ SA,9ufrG~ fS9Og쌐b3T!tιkyֳYb~Q\w2EDaبJn~OcڽҲssc Am4-%KLWsݙNut)A)ZI;TUɌKP/N}v]Z7'ݚQV RUu%ɨBjR8ͫBA?SuA^a׽Cee`JMb:d'KZͬf7κi#gGkő"'jW!_֭O=mwM1BVU)qo܄U- tcgfKƝ!-W谚w3mjt>ρfkKW^tc5d L5r^fV*<1he+wعW])[4 \|]9IzUmx\sp}y8ֽ}.m[Sթ;76&֒:jrp>r^i虊.뚭J̶*/ } wdޜ҂:_y;&J*z-e: =k"Rd:bK,M lv N\/Kpzqyd#y=,k3_jd>g_ݞBzSh0׎򼉊dYz sMC5[԰[j-n*؈IPådol:W>^dh~R+XE%$'AUbQε T8J,8MûS$@3z]O"Rva8]ӎS{v݌#_SbVf؇:Wy2Ow/ok H0,;loffeld-et-al/IMG00034.GIF100644 23423 23424 1654 6220466526 13521 0ustar luziguestGIF89a(@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,( H*\ȰÇ#JH!.Af1ǏA"TѠJGϟ ׾F{bTCjD? ڶ :~naߍ@yU~QnĠ:WJᭃ`!%B ntXK[@#! c6!_7p_aO6vzf& 5wV ׸)OIPd1>qo_h(bW(ӱюY,ÂH5ѱqhƔ񱏀a):Pr"HJ.|4P%쑘AX a(u@, V VtTIJ}^0G!_~+zG.23L&2S!4j_›C2k#9cvӜ#/9v Lh&PiZ$[vԟiH`'%(WA'(zQrӡ.(~8eЅ2EiN{gBdP<2s iUcJ\ Vz=&u5+MjupUOugU[em5=:7n20x+VOviVn~"boW ŚLͫJyUNԥլJAP~fs!iɩ&x$cdٜk2Nܭ~ҸB%vµN1r53A?C\gQE/:Q UIubή/T޷UOZ yuɺtteͿ*Z-K|/vp$\ S3zRn&krX/nZAYIqsdVb$q(L+()6-|M_zc"{~2O<&OZAs\ei9pL^rO,-*!.8)f-'sc rg &$q\m84p|LMBK*F4l,;Xˏ1=i/4#gQwMd(S[2gMZָεw^&r7b.f7VQتiyg_; og ͽT԰ݾwnz };y^N>OwdF$Qkg{cԁ5|ͥN\;݉G{e;n=t5o{cx~ݶr]X%`znڂ=~M(v 8[!Z\ n]nF"`u~e_t`6ڋz9b[qhWd(8d7ؙ-S yޖ"#8JHd:E9m83&cBn!蓌FId[iU*ziHJD盉'ܦ# kZ&aY'ZjPKj:(,J7@667+N& D ؙZ'iZcAmXie [hx?f*wcjw ѵR˵2or+קF,U ۲Htb >_6Mߝ*ޒAHtu%JXi wV.wsȖ)[4vPA2=xXk|t/N6QYYy87giܢI:V.{/o'7G/Wogw/>gt{uMԳ33_ǘ/*2a+ e]-K+҆?j=qtV@rJ}f + E 2 ;ZhdpPtrSӠxX$jl;!{hI5WBEA-lκ݈G*.%N?b'0qUd!1s+ڨw xVHJd(FL+/d"wL`pw=Py%H' ļTFNr26GՌ5huڠ*N{phbeֳؙQ4'H! `ĐQqF'EyC:a950Maթ/ 4;W)FUTRךİt lbe~N%KZͬf7z hGKҚMjWֺ%y,l"[EknQm1yftQp[8Ϲw@ЇNHOҗ;߇shPO'6q7AƳy#_iNvs2o5vZb/vg&(u?> K-Arɵ<vIP59|萩S*Eļ<&52-zٽ8|AZo]b}B'rB{W#_=&ue9wQ`t{ n_rE$j\Nlɗb h"vjcF 7l]^(#tцY/&wW@ŚiX{QHX;vxcJHh#'!%Yq3G_Uvyi"FH_.IfWag(Y*ꨤJ:hKJY*XޥrhYxRj(nYd" ٣:lM~iF%i6,kf,z*,sÊg웎vWΉe66o뭵N{ྟ%6n8Z;R+쳅R~kq!w Ryܡz[mo.\ۛḪb=[5wqqq*.HUn#~95zޣ<+>nHobo#}2l/e j\a(d7dW+7W㜇.ජs\Ξ/N$&$/o'Q@;loffeld-et-al/IMG00042.GIF100644 23423 23424 1654 6220466532 13515 0ustar luziguestGIF89a(@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,( H*\ȰÇ#JH!.Af1ǏA"TѠJGp%&yf˟2=D?q'g_u `4a.PetV7]ሯxv,g"r2(cp-8u 裄FD"PjJͅ1EkXRg`NXiemgטNFpigPi$wU3 w~sřI"jy`BjQezl25)j**J*dʹ$zꭱ:dC*JZєkUY̞jF_ԫ؎Z=-k&Sd*}0G%mG.ui߉ Z.ฃp:v잛olgë/ '~ gJ'<#lz n˻Vr;s~2E׌c6' r2,E_2H-+Q/<5+-i`Nc,u_v+aٶʹnK-Vؼ~7 Bls> TvͰ9J% +iK]1~fۏom󵘯'93;3X~bg 5SMY|Rk<|~RSϻvاڝ__*ރׯ?CZ@h 0C* Mx$|V"H)q !1Hn, ]e‘n[Tp4Ճ[pFSvxpq@*]#\x1rjZؼ%_(D!pk8FpJ_w0VMX#ұ˚bأ3bN|3qf9#"HiKzxBV"2!U"G7ʘld*%=@?Xm|,>NAŒbn;C<6әK<.s03 =.A"4iMRBS, ;6NVZyNՓ.e:PғvD]>f6Z*kS,cq(ERTe\2$ ǑӋ :QR?dhKZgִ,=޾HɍrR)OIoҍE=1*Sj}DfթR~&A5* `)_ճխ eZBMj\PBcUcdWk6jYUQT&jSNӱp]NuvZcE* [Gڳ -JE>vi}dm{3ܭsz;O+Tme{ׁ,qǹS"7Mgj]rnL;-yuk7}k!F3n{+__V7%xWzA8\da1tpz< xmg<.д-0C`N6 )nS?[;,qc\ELo[ #S8W,x{(r5"w?I /YP22S6F2s[_9<?eON3zN6]-M:y^3Tll\gN4nl,Sgt<0ivt^3]0XiR79i-N $^3S#$TDlQEcS]Wq=eF[IVij@Z؝&s9rNvMoyzh6 b)0c3kkkr۬n6p.CIyDf8qmSt:+rXKi5cJn-GPcG5́t(}>{t) E8b`ԥuΗϿds>ÙKeS{Ĉ]eW,N2Q -eլq'v?2ܯ[U݉Fv' [? ΝM8J|oetPxh6+w48WVr$3:gE78zE8T#zx34t iKQTVP$TceN2VwY8X@CNH583YgFhpmI鎶Toh[75 %o)79d5w:!zG)uycrҕ/Ie*F\5j ׈]曍׆MX~Tك: yw9{s%-%j?I(r3X8؅yJ8~@剚C"yɝ:&ȡkA_`%2i Ph.Hwf%*axbafwrsr^ Jz05fTFmZXhf#n衃.gg~f)zcY}QwhExqZ䱓޺k&2 *nR{(j+*je"7y^F)>{.FƋՔ 9ud/h+nz{z]X:hZ!'h 2swIYZ qH´2All*#cǾL_:3ϒ)'VtaOjRK`Q[#`dmhږ cu ܩh-xt+lOEjހg{rnBጣwTxWngE;loffeld-et-al/IMG00045.GIF100644 23423 23424 1722 6220466534 13516 0ustar luziguestGIF89aR@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,R H*\ȰÇ#JHŋ3jcǍ7r$bIOc-_(Dc3aH?5Ǔ5i&MYrϥ0Q3՜Pj(DO͊W`:M)dۭBْuzvQ_ūw,_T*1- #^ /ǀ!KL˘;loffeld-et-al/IMG00046.GIF100644 23423 23424 6715 6220466534 13526 0ustar luziguestGIF89a@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!, H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˗0cʜI͛8sɳϟ@ Jѣ(EʴӧPիjŪqi©zJ,ө`>6ہlUK.йv!+`Zy 6`[xǐ#,5c̹ϠCMiѧK&cЯUm6ƱmsE]fVN\(F̼d^^[s܆{qO7+ꌣv|pѧON |ݳg}AׁW_|Fm%_^BH^~)Hy'ա!'" 7,Xc-74fu Sk%…Ghd-:THB~؝EcR'[?%xbE(zQuہ'V&eqc8F&\qik>yҟ`(Ý:_{NZwةt1n$m';NIEͥf}I>+zvI+$Uny醝&嗀Jhm$l+*6h0vlRzX*ۮ+z,V(e{/2򎙬tV RY]:_˗ٱ̶0^쪓XJ\rGe>ˡ˜' ]yIg1A#8:"W,Ϋ:(jf}[UjOyq=oI!]&Y`$M^uҮ%2,;*0էڍ-,K:qvyWp!%~͠j[XZfm򨵍Lbp#ymZXIpi,TH %XG3U#i-a'Km)-Z[x&%3qr\1 Qnzҙ(4]Ij#st6N(UΙf!QItlF̰o7J$合f w7(4&+I| t,VKI3$J,:tX$Bߢ?iSNKe_))xTӢTl&ԣRl7ԪAkjRySz hs&hy$Of/ "'Ѹ>S36ַc%^}e4%IƖv֚ZIn4YegXnyc'^[ϚEy5*ηb%(Ouw ݑWzgdY/9CڇPj/6(NGq<[uU9[}db Lo1I<(>G_5|}o}E68vMg0k=~UV61y|CflR}+N&Gg?|8u_Lxc6HUiBR&kb9,xC)xӇC-hvVogok7~^o2;WEyWk7ap69C!t\؅^`b8dXfxh<dZ3rY|!mSI5N6вVv+eӲmzV.ݠ#~9v_V wh+^̸ǐ;loffeld-et-al/IMG00048.GIF100644 23423 23424 6317 6220466534 13526 0ustar luziguestGIF89a@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!, H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˗0cʜI͛8sɳϟ@=ѣHҧP JU*P ժׯ.2 حSj=K۷+µt-Zx˷X}_VKn+*wǐL˘3k̹ϠC_,$BQO{5Ӯ5<;cBk㦨sowo .߼H|A={نKʵku!V/^|sex֞V^n>_VߵwA! ǐ|N`zA nZ|a(b :s=E4^((݆\G"$d]9b DʭTm>cECȤ6#aEZWeWeoeV ^~&nhʼn'sW\wsU 7]Dvt=YɝJ嚎N$$e&ݩ~y}(/r*ۮ~+2!H뭶槬}w"VZYxJenx*͊-],r켐.['v+,,!|W;pAWRj꧓BpJy˥?krnXp2 hkk%u:)vύ^4'Q{ڡ49irFѱG.^ipF'jEfK:ʑ'=\xG4[ s[Rt<6`^hH w j&S[dbR`㛔DES~%KFRNBR/hKakDc&% Y#YI8jf"e)%bFrE!@8LJ훩ڄ)7IT7xe&ꗢdT3Tzљ\Y]sr7%d T] (Ⓑ0|b,!75PTӘܙkѕ>q3BB2aif\LcZKys\ÖOsԩJ'Oz>Z47wXM6r$Q[&\ҭZlc5褑2:,>LI2uY[M[j;%hXaM̯@l!KNJ{UU.>j hGKҚv}=-_JӇ W[W@ͶcltrK[ZUV{K}ˊC.qk2劶 mpKZͮvz xKMz]3[ݦvƙnb _1W+L^|f  `1qp_,aVh;8T' 5qts:PGZpA4=B%h.;6[7%CTmu@lKZa0hKu{UBP[ԠfݵoM<6+?,y~,ċh G$1F\ڍ3*=t2?d_C%WEՖh9Pm4tER{}U ڣ`u=iG#*cQ!+Y "Nz73@q+mm%+𣯆c0Wzzvdfgw;^eoB،3pqO9XOw^=claG~@ ~9i<]ϸJx ?. cN8Ϲw@d7s*Bё.􉇉ϙW7uL,m֛ŏ B:<NN}~phDnkrF) 7yk=5!V;l&;\ƥSܟnkc9][;Vuhrn̆KQU⭶:VzKPXǪ39\,ѵElh! ֧B)'iۏ|3~qRw[u[>r ]d.Ei˵S^,>;wgĆ[OUlУKN9;loffeld-et-al/IMG00050.GIF100644 23423 23424 1656 6220466536 13522 0ustar luziguestGIF89a)@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,) H*\ȰÇ#JHQ".Ө0G 9 ȍO˃*c.li͐iL r'ό3kLK8I)Т eԤQK^uG:Zz4,RM*"I=;Hݻx;loffeld-et-al/IMG00051.GIF100644 23423 23424 2070 6220466536 13512 0ustar luziguestGIF89a|@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,| H*\ȰÇ#JHŋ3jȱǏ C$M>4ɒ“ ab˖/ Tٰ?)!O> zTΠd*ԞPR͚@_ڼXQæsIJ]j:l׺sz[~.} /aŋ!?=SbYϣ) vUsP| w˺uNFiegr]m2~澍pα{T57^-Ӧsxjmʽo˫_Ͼ;loffeld-et-al/IMG00052.GIF100644 23423 23424 1553 6220466536 13520 0ustar luziguestGIF89a@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,H H*\Ȱ! @/VX1Ĉ?ntѣ’Q y"K9Z4p̘2irɳg€;loffeld-et-al/IMG00053.GIF100644 23423 23424 1501 6220453620 13500 0ustar luziguestGIF89a @ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,  H*\ȰÇ#JHŋ ;loffeld-et-al/IMG00054.GIF100644 23423 23424 2655 6220466536 13526 0ustar luziguestGIF89a6@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,6 H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˗0c fE-qYPE<.zPF!&=TROʐTV;f*`:},Afi,԰ FTvkYTO vwlߞu[c xmqF^yO-{x3hŜ >Ģ?cLl]dW{Vu޷W0ܶ([Ε;mrQ{܆Acq7֒9 FRN I_frlhIg 5s7 d%i> &{yczI*V'a&!|fljfL:7`]jܨId8c*F'+RvOJ^:flPibQv*ɬ-ZJnf:]j(Bz NTH]eErmHҪ/JRT  #Wlgw T(AFg"ݹiMTPjI*ԬG"We%cεɶE&Q ,`ӱa+K0W26{rd˓=g/vF|4Ӥim7n¿M_Nrt]Mu:5jŘonÉk[|#Z3VN\r&K5үY~u)>]-{M~ݺG{_gaހ_pŇ z}ͅ]w2hsζa|UH^PXYtcU"ɶh7`yw!ku]D+"{?taJȢpC*9G=`gwɨ!NU}]$+Yҩq֦U$]eREb|J9QB: !ZO)jhOjΖ[VWPbgժ7:kQ롿kk&6F+Vkfv+k@ +LkoX]/I\Ưzkpeua ci6,qxZgLqe\q",0,4l8<@-D\0Gߜ4EO*3Ctnj,1]2ћe Lbg䥄c;Guq5ѮZ{RVxmn#] ږd˃pyp1'cV9X ,8u_Ic +k;^Q㭣ltj*.rCw8 +rCgClBT> #? -jwaOjg͉ˎ[K 8՟4fykr+_ Z̠7;loffeld-et-al/IMG00057.GIF100644 23423 23424 1663 6220466540 13522 0ustar luziguestGIF89a,@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,, H*\ȰÇ ańfp#ǂHGG.Q, %J'[HR"˃:Cڌ)q6u&K2K6Ҩ*MR̚Ѭ+1"֨d1yӰce;w-Ժx˷܀;loffeld-et-al/IMG00058.GIF100644 23423 23424 1663 6220466540 13523 0ustar luziguestGIF89a,@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,, H*\ȰÇ ańfp#ǂHGG.Q, %J'[HR"˃:Cڌ)q6u&K2K6Ҩ*MR̚Ѭ+1"֨d1yӰce;w-Ժx˷܀;loffeld-et-al/IMG00059.GIF100644 23423 23424 3051 6220466540 13515 0ustar luziguestGIF89aI@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,I H*\ȰÇ#JHŋ3jȱǏ CIɓ(SV@˗0cFl)͛.i* &O,q "Ye蓨ӧl)uѪS b֨I"٘_Z:vTj6ֶknO?үGx%̘.ӿ; nLabƭ9*]tsME͚jȭc Zm˷s~7Q޾.KƓDjo2N=tDZkZ2بEOV d׆twwn$}toEW͗}e~E=fivP{!6o'bJdj؄wby(L)ވ`ňX-cd^55xEw^fYek-VQJ$!n6 8%9i32xJtb$yv(L8~ ^a6&ɦqy&2Jъzv^Fb}]\:'{Py*2SajzP.*mެRYyv٢l"[h;cO:*ߴx""Ɋ;㼀 Zg)]v?W?" 7 o{Rlg,˱siy",q% k&qV<-߻e!YsN/,ҝ^zlBD4t=mUQ N]WD*̨;r^o 86K5Y626C=wr߭ |2;loffeld-et-al/IMG00060.GIF100644 23423 23424 2332 6220466542 13510 0ustar luziguestGIF89a@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!, H*\ȰÇ#JHŋ3jȱǏ CIɓ\2ʖ)aʜyL4sޤS'O=Ѣ"|iҘ>tTW7̺)VInp(NcN1-E]6tv݇tEyb߲wXM<^2*9fհ5R\reĠ5K}:1̐z ͬ U u뼱Cפyw6͖.8k{En{pֱKu>sO]<筵 Mx5v9\ַc-񹖛O:ໍRt {MF^y)` Ytmfn._v%~jxlQx}"Fۋ~ P k2c-X[WUXV7Q3ti'BF>6^cUOb)ZfGhp)tYR@;loffeld-et-al/IMG00061.GIF100644 23423 23424 2341 6220466542 13511 0ustar luziguestGIF89a.@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,. H*\ȰÇ#JHŋ3jȱǏ '9"I!STd˕0c83&ƚ6drgA9# JhB OJ hթ6JQG-Z*+ͬT͒zU+صB Zsֵ+סYNw(hړ=ݽQE2KgQ]U\˘jnX%BħS#}L3lOC:viN?V{im92ݮu_Lx8q>ߦsoGyx9G-5} >QgVCWrg!ew_}&'wnzԃGևZ#w0hdEh^u7n\If#'`A؁&ף'.#1iuЅZ8aHBx܄jem._Wsix;loffeld-et-al/IMG00062.GIF100644 23423 23424 2725 6220466542 13520 0ustar luziguestGIF89aN@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,N H*\ȰÇ#JHŋ3jȱǏ CI$EL\ɲ`ʖ0c|)͌4%ɳ·(ysĔD}Ʀ $RӓF+h.k-Haw6jVVh_5UHڻ2j-lIS/{re91rcZtiGڢ`Эc#Z6em,X3pqE פ5e*Նk^Žop[#}tY9; Z]mGN}zq!w.fV؜Q>!1fe$WX]^9u1b ҖXxxH8ހxH63|GLV(tXfIY$Iiޑ>bI_]xݙE f7&QY}2`H+gY&SP!y8tieb~,)Y5RtJ8_Y^9"|~)gbZ%Ěuڊx k[kkk.i"@>[-DzD䶷)6-u.Kٷxn`GkoThc9*-kZ5\ \!p_޻>lhGNޫc{0K r%LL8e-7운?.滑IpoeK>9Sb_֭C3a};WMo5o{n˟lq ]˯W:umrAWk XYmx!hqi֑{y5uV-8]|9MZbe&"9w\!E&F"iQ#)Jx҆HbbF"hBTdJGӎ;%PSFdXf\v4;loffeld-et-al/IMG00065.GIF100644 23423 23424 7554 6220466544 13532 0ustar luziguestGIF89a@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!, H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˗0cʜI͛8sɳϟ@ J'E ҦBj)UTgbp+ׯ]E %ٲ Ϣ]T,QlQۖnMn߽Y)0†F;h3 v[ɔbj9ϠC&q&>걬[7|lScN6 q߽W7;B#X`Oepg<o#·t$M2Z ¢k׷}+_vh!HyUh԰*|J:DK lC7rP$N7If2Ҵi?7xγ9=sh>dK F"&zwcjj0*tshHٝCGMRԨNWVհgMZָεlbkʶ*<H]a,1tu[y/jXZm1;pv= ׶:MT[aK^.l+ZNŽM^vt7U ӫQvhYMz80~pC'l.xgzoP"<1&doM/ݭuYq%Oڄo5w{bǨS,!縰).M[l+f&-ߍF~ͻxo̵MS~گ$Ctn}e]aw FdIzx^aH߈CxYX!(]*bƢq@>8^B((>k76Hdrz"sƜ(&d[ڙvXދރm6oۜGާbezR' xǣq&Z|)\b-Wn'\**h~AZ"r (KIZڪH*+PV~l&&$*MudrjA;epemz$-FKl妫nL㶇ƋRh+J,l' 7G,Wlgwޚ r[((r/ p,s6|;o3?/tCZ49#}KwI_N]T5Ysݵ_?cdmvj;ͰpGwםZwp̃=Y`-.N_Zs./{vqZ ݸl 쏶YϦiy%ϵۺCBg]N)<#Ojȝn{ힲׂn<}ߪ}w㋏"KnH(c?fzUڳ޾Qz}\󰓦]w,puFt;On\]/}c 7 _|#ԉw6!vB 8)B x9uqMxp1/Y tG#ZRb`$|ghK \=ٕ1xƓ") dX(ґ/NF$(@p{|-jVѱRf XEUִa}[Osj\{Y5h٫ZۊWbZ+l **֯}d'KZͬf7z hGKҚMjRV~mKhK _qDPp\,vȍ^rK@%]P:X^JsKZwmw d *BK^vϥOCJ-&CY-|{!xw. )NB4O3<Ɂ̡ _Ϣ&* S1rp؍ûnayGhF"eVI*˄dt=92rll$KPK}H'tK`}: NWeϝ{=2ʼR, [S VC^Ά9ST9D(1e=1a Hnd#+ł} n{֜҆Bf:,4i&KsT*P[OOv\(PIakϱbF]mc^";)۰^- EARm4 0lQCaF(D5_O1q vj#üU-a^c3:L :ƌh;BNPx~OՖN͞{BNktwW P6M۟|u5Ϊ&dj!~obkAz::mnyzee»o3vvLw#> L}M<7Ou{GO;loffeld-et-al/IMG00067.GIF100644 23423 23424 2545 6220466544 13527 0ustar luziguestGIF89a.@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,. H*\ȰÇ#JHŋ3jȱǏ CIIRǤhJ9{-d$]Yw)Hٗ e] 6UMxb&!Flvicvp҉'o]>Fgs~:aZHFYc3vi~XZb]W bU(뤢vDF y'uijR:!Bheaa"zlqXEgC+#bkokV(l^~zpz؝ ,RX‡q|dZj.n>;ܓiF,qh7,KR(8t"RةT?lf⥁unUC2*_]{,ԫg;ݿzt/Keiih( mcͷ~j4e^G/ؚ ),xm[ӚgyQ-ԨzLTʶy'< kWo w8{k硪jso>M}UoT*8J>~'lxY(;t&-REZ; 7h= r GH(L W0 gH8̡wWWT!wÂ8%LԵ'0A3dA+N%NM5ɋkx S͈xbx2] 뷝.1 |׏@"IRTݛd=rTd6"^26$4Mk*'Jre*m7".52iD٣ڔJ!ɐN-5W>΃o%05,Y.#5z[04YɑfOR0c+;tǩXz$ҒC:yiEӕ<ђ3R#N3% uyRR+OE4;(EYNo>IT+i"1eK);cf1dN2(@2͜ڼbLE^A-KbEF/>e[T9BmևYɭReXi!C3kUB8X5ؼ'Fj:5T,*_#aO}`aX&WcX˼N͓e8iuK^ w]tj2E%qk29۹==FcOU\Vlns*R-vz xKMz^P PV &^̉] |[zӰcaƚ푋^ֶz۴&4\1{3%n삣ji.~#"JrpXqdjZәzxYcUJ& 5n_Vزc,@:>X@\Tzs2kfJd ]Nvv<.I]M\:N/@I9neޡqš6gf4X,_imNm-h4F=f6@gX^,ǶNB|m60kWЪ5JN~ʖWUd2elJk1Dja(Hjg#[[a77MpzO][ܮ[\vy[˲&.n.uMmW9Ŭv.riK-[S/lAav6\/WRGk_&?[d)M[nU~鴊fpqg[l -j7D?27jsCd— ?UךcW.YץrηuMW?ثVϮr=ra^s{u17or_q"Sf.aЛOWֻgO={NW}ؗ^KgM}%'trմyď&Xϸ/?fnFw%Iovx?~-uqwb'iBJ6CV;{ӶDSVYaYՀ7w3h8y(ws${6?4!}qx"G^&9>K6<[aj¦~ XhCu(uwoUose"ERXDҖhEÄUPI\Hl7wVozk0EwS($_Nd4uqvT:x7c(0Mׁqmw'&\74skXr7uoՃw-0u{X~Hȉv(d0,ȊZLE8lhSË>9|Ǖysv>Ut?׌&xؘڸ؍8Xx%؊RiԎ=Xx;LjXwo@(8r^HVwWtgDhE҇Nޣt$G?()^(s!0N&h~Ȍx (4]RRƵOZ0Ff÷Vt28$z2C,x?9[B822&)p9WTXa(]PX-^d )XW8sy%c9tՄ9-0?*66~^iu]sqq18^|2' W>3KR}/E{χQ6ѧdLSdTyG6F+}=>cS{y881|OW8wyɏșʹٜW$` }qbynIwڹ/Y9YI;loffeld-et-al/IMG00069.GIF100644 23423 23424 2672 6220466546 13534 0ustar luziguestGIF89a2@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,2 H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\R%-cʜIS 7k͗@{%їBҤƔ 'UEVxueמZ-iuƯ)т5UI;zӦdʭW)ҹ*Ԋ.Ae&7㼁!_̱焚F 4柇q.]w3FӪ6n3lƦi{d⧟9+۱naw$=\$αzمL1s[ O,\6\QoqG] X ~z |盁gWa>Wp>߉(^X|W-Uc6P(y,vf!E (Jnmd(Wv_eځV>f(ٵW"f9a*R F\ujؤqVg5'ʔGm~%inr‰mc=Z7gz SpE:%jP9jbf5NJjULx[L7ZZcڷaG֮|%g Kk2ꭩc1kn%Img'B!>:obiԿ 79,ĥJY&w ,$l;loffeld-et-al/IMG00070.GIF100644 23423 23424 4646 6220466546 13527 0ustar luziguestGIF89a@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!, H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˗0cʜIs#oɳϟnJQ: <:Q(ӧPq: 5ҪX\z5k®^*Xg*UAn㚼T@v~:.^5+pെKֽ 8[:~,آξOIW!f˔nF81;\s/85dJ ;:mɖl껆^qSi*oT=SCǹ]]|t᧛{֟;41y'^d`qV`ZGt `ǝf6Gn`$R\|ZXy% [T[) 2H,慣z9=hVQ/z&lZdjLgړ7eKCܓ 'ܖ.u2Zygo&Kmr&Icw'R(*'Ri:L_.dfBM>:bi)z gM֢ \zun)"bbƖjk^(JylT)Z驘jVJY묢ѵƙ~Q,{uENoXhK\+plv^9i0undFp K',s[wWf<1\{?ʜ]he#S%I`#byUU/tX6:ƳX=R/B{b[6׃I{6cw6Zʞ(G5[y GO8 mrƭw9#A9H~~C~4H37nsilrW9oU߲{֮I=ʇ|Jٯ i!}+'yt'vM@e~q7;ӑҕhr`H2:s'A ⭂̠7z GH(L eT:.T c819\ĵ&Sݐ2rt";ݚm+lpDmsbe&V([b ".0bL6pH:x̣>VdX3ZkHW|ܣx"QLIuӂFudW!o]+!&,F1Z MJ;t/Jj&Kd9e4zkbAө"StuCUP u4wTMԛd&E+1sx]H;}21բEeTjicUйW'{MhRQׁ+ƴ1jjcM')`8Zk]U'qx-QA4M,T+\3Wt6\>Qv]+S}d_YNoٻXioc@B3e&k^Lwn\|2"VofHOkqIo.V.a;loffeld-et-al/IMG00071.GIF100644 23423 23424 1716 6220466546 13523 0ustar luziguestGIF89a:@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,: H*\ȰÇ#JHBbL!ƎD8dȂ%M.LiA"Ȅ qL%iҌSO:O̹EC_&zOOyQQarUN4IeQX=r;dϱdh/k̹Ā;loffeld-et-al/IMG00073.GIF100644 23423 23424 3320 6220466546 13516 0ustar luziguestGIF89aa@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,a H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\r%/8Sc1o\s'ȗ=)%8"Ѝ:(iȢANf!֮6LiYJ,lԴbe^ũb[p=k#٣Zk0žsR 󝋴pN%-2OύNkcQKO:0Qω9άNjVvwlϦ;WԀO lM͇WC] "^:eiv<4b=~u/\F9fYP}[~^ui\{ qtށk1h!sywbۆ 4H)7^.^hZ'cJ-T7nR|W^B62uNVI}f%oVv`)dihlp)tix&z~j$9eX8^[F&'`9]\uu(i?V$d$") *()n) %E莱⬶c+k&6F+Vkf&-~;~ڹo dfkb%r+oF{ﻭ%Cf]+'wp}~1\>!̟v ^c5':1li!ƪttU ;ڔ-i9 ۢ3ljjJIWﯵTdh/k̹Ā;loffeld-et-al/IMG00075.GIF100644 23423 23424 2353 6220466550 13520 0ustar luziguestGIF89a@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!, H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ%Aa@F#q.Y΋:yLsO5ygcӃIiJ,gT N4֥$j_#ZEdځkRػPeUͽyUPpekxgYq_E\7ڸV5xCo-s2bEZ,瘮?[|adJ͛hM&mлW\QXk̷O=9]|3Jz#]~4}T}߸}[kI]vހڂ1X^hXlgaXXQGo-[jJ&a+(MvU55[gT9G6$HE^7RO$R^%ZtʕaKPNS\%G^YTl>&lozhetix|矀*(J;loffeld-et-al/IMG00076.GIF100644 23423 23424 2525 6220466550 13522 0ustar luziguestGIF89a)@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,) H*\ȰÇ#JHŋ3jȱǏ CIɓ \ɲK*_ʜI&5sQ%N@ e翛Ho]T&Χ>mJjȟFZʵgRvK6#VeӪx۵p|Vܻ^ԋv)ߞ.8!oualc$1A-O;2G̋?7ԌEӝFX4єIauWiϮ 6olu,{wcoD]8۶GX8̉S7򢫧fGUlXC t#~tF\xM'o-g~Yv^G `7ၯ__b v$WɧfPBu2NA|J-Wb>vލc{[]taa)at#eH( f>Iz, ^g9)$Iu O_y_ʝ+YYcٝHvbzF⤟ǓRz)y^)46p)GN|rƓˎ}c_霡ף^pיλws\zb+g;wלkƕ\x Zԟ|ѧMUEՄiރ U`m# hΕ+`J by >"lyGP xb}[n5zZQw%(E`_FGǤpZq^9!MellŸfN"GgQ*7h疌&fu_bga&wԒ(F]JGz6q*IGBe!ɹއgY!GJ)K5H,{EI5!)O lLSNVj=ZWu]-f:ԯۘ0'slZn Ue7,qBcBdtj:g Xp 8.WgYV噪i3^znRफ~xNJ&sZfj"爻JJTRIw8ڦVQ@Ei֖(+k覫;loffeld-et-al/IMG00079.GIF100644 23423 23424 4274 6220466552 13532 0ustar luziguestGIF89av@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,v H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˗0Q pA1aܙN=[YfPG%MzSK60*UOyUfW_IMh5Dzf=;ѵg]%:W+Tx*D;)WhlVpɹ Lq^bKVر/.e?CsHQ/L:[-IWfG_!ve|}뛩?ǝ֜ZFm}k)(`QEh}Yx]ሷmXVmq"ߋQI#(jفH!5b($>x FcuNqKXR$g V>aOh+n1i X*%eY`yjoict!Yd[>rJIh`Ib c7k̍%|`jjxbJ^F{^_땉J⮶yxWGz*[kQ螔fGSʖulZl%n:nU[b騒e TT!zd{ȵ뮵6wMu(%,w!\ȵ91Wĥ+p$2VZ+,lɄj,ʜ%,nl3<4a `[=4W}y_MmXg\w`-dmhlpot[.7GN}7×(߄'js7~ 6A~ JSGuߚ;騧ꬷ.n/K2TI vǖa=؇Yϔywm4G= Kr8H!e0`o\㹾)g߽ sĀaJؿl \{7>=obRQfʂX1nY kcvbHMaġ Ati\8{vB8zI}03 '4^p_PWa)Ë-!7%!-aWܢXȁSQ ',K"TN7W'?Yo6l C*r,Y"/ZʣxiyɊr$$GƘr$!(P Q`WLC::t4}2),haJ4Ote>Yt2ʾD*]v[I΍ה ɗXM81guԦM2jj gtoߞ'>*\On Ӆ7<`&Mȣ'6uNE#j\C!ʮk^46t- Hz ť|# =4!?2qѻ(jQM.*y}Eu3y6`~ ᖜ)jERG`A w *}]S6q8!G~ ={/"Xщ7^ݍ6X*f0bAh#Q>RHFT1rTVQ%hЇIrbCU69fEő %'`$ejDtjVGDtq롹bTq駡?bx [fT&%X)h6JE5ْHm]kzCJun:'vitSjl>IepneVlZmخ ):+aM9~fGkǟzTj&B9縭k+e*+O *~:p *?mB qʃ n9\r?2IѺ:-̭|0m. ȈF茿}[p^1mUm+J4樲Lܰ9i o,jipM^,pQs7*>g|8eK~n59< _ Nz433.S+_;b_eoG#y1'{݃oNFm}?6ϿoV߯ 7̗&w% HAH̠7hr}GD c-z gH8̡w@ H"HL,.CX'QqЉ\|Y>0Y,/5ގx/rc`tʀxiu;paG6r.ߠMCa H݆qȻJbأNX'a&g1"+ʍZvKCQV2ڑX2<^BjV3#Cb,LC3fE]bK 3cx,^w2*dJ޾y5_nc@{ '7s ] +S:*<6DE :M 3 R cHRfyH/JG > tk:(>M7% M6S&WVdc)i1 k):RV26XVSCG>*%DAVWsIk7Fo^ m9메`IWZo m }sWD\ET4KQJ6I]Q[t#jmf_%ׁ4"G5V :CvqӴr^ X|%|6~&hp \eWEhc>"\״>U%AvuZ_SPFj`CS'V|ejiCBQzRr\#Fm^/pFmZwh3gMOXQ3, ]%S/?w5Nbdo 2^WUm0TCYg1N5qgP& RmmLRYX+)WV5iT74r7oO5o/`xD!L(s@U%<I薉f8x<3EW7nQi} |N(ѐyu5X~u&g)K&)$F -k`}(~ᄒCtܢb%UE.9z-@Gl=8+(zCzj[\ٕ^`b9dYfyh!jRmOo)`O!At e4A3M9~u vQ3sɘ_yٙ9YyI/}),ĕC`^ӇtUb҇{Y4KIl%QRɊǖJm&) 8Suej FB al[ V?I{o9cqESقIV)O 8H85HJ؞BhxrH)X(89/K~7uEURpDVidm`؂؜uIi3~ERxSAs1Xȏɢ>Ꝯ{.nBgHQh/ڇjIhu8XZ7S{5\ 9k2~ئ lRhpئJc8S< *)n7Kؒ$(uىܦo5lMgovV#/Ȥc7IPN6xax2cqT8He$jV䄌ڥhtdcHr&)o* ٫:jeJ/ҭ ahvfi繑X Jh˓Û,y6wcn~%ĒuWF~8yiD;-tsH{\UIg[a% qzUo"+<۳>@B;D[F{HJL۴NPR;T[V{XZ\۵^`b;d[f{hA2i+g~ïضHdZrkjbz*7;Ŷ| OkN}5Fb8nViENz[5 RfDW5ndK/ǹKi[{ۻ;[{țʻۼ;Ի;loffeld-et-al/IMG00081.GIF100644 23423 23424 2655 6220466552 13524 0ustar luziguestGIF89a6@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,6 H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˗0c fE-qYPE<.zPF!&=TROʐTV;f*`:},Afi,԰ FTvkYTO vwlߞu[c xmqF^yO-{x3hŜ >Ģ?cLl]dW{Vu޷W0ܶ([Ε;mrQ{܆Acq7֒9 FRN I_frlhIg 5s7 d%i> &{yczI*V'a&!|fljfL:7`]jܨId8c*F'+RvOJ^:flPibQv*ɬ-ZJnf:]j(Bz NTH]eErmHҪ/JRT  #Wlgw鋃+*>~C%Nb6LQԜG#x:V y\֍/^׫]׶"wLk/Cd{]O߯8k=>m>ﯝz}Ȕ׭ƅ7K,-#I 5Q9 &*hFy ߯6V!< !H;)AotŒ5zlo C G0 s %({C՛8htDO@/;!ò>?<ҐujQ"Pz9.mrsQtSJ pw< H^1g!>W1PC,D#HF[sȽ&Xk h%]\ΦN)eЀ0+zAX} "mYfCeB DRH|W%L3x23xyIB9i8SAK6s< _ɳZ>d4@xhV;}nZ#9Fws9"1o6„sy2f tf_ i66y."5*#A)uHɗ02iHh~5|LHP;h})Q2r,P[4pWBhN6A6L:xγ> JXB?\/q-"<`U}Pʞf IthԨNWVհuXY̍]me߄nBWwJVf65+Rkq9f({:'ζn{MrfH@;loffeld-et-al/IMG00083.GIF100644 23423 23424 5466 6220466554 13533 0ustar luziguestGIF89a @ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,  H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˗0cʜI͛8sɳϟ@`QG L(SDN|*RZ3UXnm8Ver]4aڴbk>ː.ۖEپnD^q).ްm!w.Ҿ,Xp㰍v>WaӏKk<2⨣%3^(t^،džX9VWo鿨^͵97 z4RN>8vnp_ϯǻ(hU% 6F(VhZWz!}}q̉Fbsdj[#ġS`-ʔ`w ј"~ 霊uIY҂ڄ\u^ue^{9~%\b.HecdN )ׇ!IwE򔦛4e^qZw0aONUx&ʦ6裐X|7h0U馜 ik]:f jWGJWKr%P36'tBMwgڈFr^/} 6:Y dz}7Gi ám8j,xϽέfZ`JK=k|zgFy:FmɨY6~$ͺnm#]qkī9&Ef$֎X'X4fE4IZbKnH ~w,htBQQpLGUjʈGdj47@vX䴺FӜ'YeKb&,Tɕ!DAXJ d|M*2Ћ{;-ո~c[`N 'ޘJNms/# 3[d%]yKoR[c6ɑ sp:,m vC=Rr|޶o2 QBX@".]pK[!O UE7G0ia' Lsg#%W R}\#9Pr/XfҞoT*Ltfq8gT 7f&g[K 'L~ KS8v~j&M0MXB(*ѩٳjѡ0ZLd'FcZH ޔ?K$,5˛hv~CTbP2 PhPvJMuDvM* ,2 6ζ0sG1.X_R[u2W\Rmc2עh9-5&uv>1tGW="3=eӜo9cєRYl `YWVLK/[*K7/wN\-"5(msBtm7 馪t-'%T}R$*|cXZ[,y8mbO9SʈUjX9\iZͨr[V؎ΥBy{څO5]]k $Dǒ3rwv38%lw^U [όUr\0W;jsL~eʵw^ݬMbNf;ЎvxV}6SLnx({oN~sWUZ 7YJ^[v&zη~Ofw?໼}y'&ڇK˸ޟ39*޸rEn(OW0g;loffeld-et-al/IMG00084.GIF100644 23423 23424 10341 6220466554 13540 0ustar luziguestGIF89a4@ @ @@@@@``@``@@@@ @ @ @ @@ @ @ ` `@ ` ` @ @ @ @ @@@@@@ @ @@ @ @@@@@@@@@@`@`@@`@`@@@@@@@@@@@@@@@@@@@@``@``` ` @` ` `@`@@`@`@````@``````@````@````@````@``@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@@ @ @@@@@``@``@@@@!,4 H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˗0cʜI͛8sɳϟ@ JѣH*]ʴӧPJJիXjʵׯ`KYҦ-6ڶǪT]rՋm޼&|NU$Ɔ#D/ B69qǞweΠ7׍ 5ԒcO\2l.ymnڡU..췗K=\:龹;?;ч7;C=|n>4}ho =&^&Wy ߃Cz{%`}ՆGrU wbuRa,(4h8<@)DiH&L6|PF)TViXf\v`)&]1כD g jfnR`Dkax9vRty+"Z}f(ive^j֟%#שٓO#D h)SMPnH[z9߰P(La)}B;Rp=ֹ+~Vmۡ ȩ;m_n˞k~]jT^H~MΊTg J_pəufrvZ K2r2ۮ3s9V&3[<-R xL%Q+G;* 5sL;5ڷn2̥[t[d2~u)keJ8IR#/W=kr 8$.nfޠ葙Lb(?&3w41") ~"z:Mj7U~5o :oR4 z=)(Jk벎ץی_%'I'\)ئqټCu9FɌTyi$:CMPQeqo '6AӪ*'votLL*[JўS-LDh-u +Ϊ&̙',D2\F#fO f3#wzCZl V0*p]b |bdyy (YQ*WM+hCt󓘭gi{46okI"9+`ۊ햷k_ YSsҽdKj&M_̳BlۨMZߎ־!ma#Vfլ'hkM GWg;KɃQk"Evl$*zZ!D.l-mnq ԟ}u6~7ʸiGsv$ewcmTOj֪Uv|c'QB9f3*>'y9q0}w7.K$O8_ mi }$/M[@ I>]g^ԁDȇ[exVrcom8腅xtd\\usXd;Jt77u2Ihq؋gWaS@xIXl(z)hyf:W{eT~ +=4(LuutC\~(-Ŏkxid8([bD^vzHl)oXk~]199WPؐv>,F|CnP؊.wW8f1=r;B_;FV&Ibވn&YDŽHS% (h;&lЇuPVh"wB+H)⣗g% vhpLw(r$ טFUL5mIAd9p_Or+H3yGe!9i2阱HfY 4GQCWH$ 8Hʙhu~JɋƉc9胗d!F>xe~ -Yx)5tHOkؓ t%wH:s8&:Zz*%Iz ":$Z&z(*,ڢ.02:4lr3BQ5_lu8&V&:eA9'ғ^ItZ88b/Y{AOh\9YGe6QZ]:d*z)\Z)釘&=hzWؗZT9ӑ_NӴn[1`--4Zzڪ:Zz'gxa04m]WY+*.C )*zX_?  .9SZcΡ*:Zzگ;[{ ;loffeld-et-al/loffeld-et-al.doc100644 23423 23424 2176000 6220453256 15345 0ustar luziguestࡱ>  ZWX|Root Entry F=Φ(MΦ[@OWordDocumentnObjectPoolC$`p=Φ`p=ΦSummaryInformation(Y  !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOvSTUVG]\^_`abcdefghijklmnopqxyz{|}~CompObj?B% _ObjInfo! Equation Native A" CompObj; j                     # $ & ( ) * + , - . / 0 1 2 3 4 5 6 7 9 <      # !"$%&'()*+,-./012=456789:;<?@ABCDEIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmopqrstuwxyz{RLq$  FMicrosoft Word Document MSWordDocWord.Document.69qiscrete derivative and integrating again. This procedure, conceptually appealing as it may appear, however, yields strongly biased phase derivatives and thus strongly biased phase estimates as the paper will show. This bias is a serious drawback of any differentiation of modulo mapped noise contaminated functions. The problem lies in the methodology rather than in the wrap around effect, as no sequence of linear and nonlinear operations may be altered without seriously affecting the result. It can be shown mathematically, that computing the discrete derivative of noisy modulo-2( mapped phase yields estimates of the unambiguous discrete derivative, which are always biased towards lower absolute values. Thus phase slopes are always underestimated. It can be shown further that the bias clearly depends on the phase slope itself as well as on the coherence. The paper will present the theoretical analysis, and give some hints of how to circumvent the problem. Keywords: Phase unwrapping techniques, phase slope estimation, biased phase estimates Introduction The determination of the unambiguous phase from noisy observations of complex angularly modulated signals is an unsolved problem in general, especially if phase and amplitude are mutually uncorrelated or even independent. This is clearly the case for a complex SAR interferogram. In terms of signal theory a SAR interferogram can be considered as a complex, simultaneously amplitude and phase modulated 2D signal with non Gaussian error statistics. Usually the wanted interferometric phase is obtained by a simple tan-1 operation delivering phase values within the principal interval (e.g. -SYMBOL 112 \f "Symbol", SYMBOL 112 \f "Symbol", depending on the kind of inverse function). These phases contain all the information needed for the generation of digital terrain elevation maps of observed areas but they do not contain that information in an unambiguous way as any absolute phase offset (an integer multiple of 2SYMBOL 112 \f "Symbol") is lost. Furthermore they are subject to phase noise coming from the superimposed amplitude noise in real and imaginary part of the InSAR image. The process of resolving these phase ambiguities is usually called phase unwrapping which in terms of signal theory is simply a two dimensional phase demodulation problem. Nearly all classical approaches to phase unwrapping, known from optical interferometry apply a sequence of differentiating, taking the principal value of the discrete derivative and integrating again along specified paths. The paper will show that such a sequence of operations always yields more or less badly biased estimates of the wanted derivative of the unambiguous phase. This especially applies to a combination with Linear Least Squares techniques which are commonly used to reduce stochastic phase errors. The paper will present the analysis, the numerical, and the analytical evaluation of the phase bias depending on the phase slope and on the signal to noise ratio. Finally some hints to circumvent the problem will be given. 1. Theory and Concepts 1.1 1D Stochastic Analysis Let the observed phase obtained from a tan-1-operation form real and imaginary part of the interferogram, be related to the unambiguous phase by the following mapping:  EMBED Equation.2  ( AUTONUMLGL ) where ((k) is the true unambiguous phase at time or point k,  EMBED Equation.2  is the phase error and the bracket indicates the operation of taking the principal value of the argument phase term in a way that:  EMBED Equation.2  ( AUTONUMLGL ) As a result of equations 1 and 2 the observed phase always lies within the base interval (-(,((. This effect, arising anytime when the interferometric phase is computed by a tan-1(() operation from quadrature and inphase component of an interferogram, is well known as the wrap around effect. Nearly all known phase unwrapping techniques with the exception of /LoKr1, LoKr2/ try to unwrap the mapped phases by a sequence of differentiating, taking the principal value of the discrete derivative and integrating again. The operation of forming the discrete derivative yields:  EMBED Equation.2  ( AUTONUMLGL ) In the sequence of operations in equation 3 we have only used the fact that adding any integer multiple of 2( does not change the result of a modulo-2( operation.  EMBED Equation.2  are the phase errors at points k and k+1, respectively, mapped into the base interval (-(,((. The stochastic properties of these errors, namely distribution density and second order moments are known. Now again using the same identities in equation 3 as before we can further write:  EMBED Equation.2  ( AUTONUMLGL ) where the last equality has exploited that along normal integration paths the true phase variations modulus is always smaller than (. Normal refers to those integration paths which do not cross a cutline.  EMBED Equation.2  is the true discrete phase derivative. Equation 4 clearly expresses the error which we commit when forming the discrete derivative from modulo-2( mapped noisy data. If there were no phase error present the result would be totally correct, but since phase errors always occur in normal interferograms, we commit a systematic error when differentiating modulo-2( mapped data. The phase derivative will be - depending on the noise - more or less badly biased and so will be the unwrapped phase if the biased derivative is integrated again. The further investigation of the bias error will be organized as follows. In a first step we will investigate the stochastic features of the phase error difference in the second term of equation 4. Then we will evaluate, how a modulo operation, which occurs twice, changes the known distribution density of random variable. Let us introduce the non-mapped phase error difference variable by:  EMBED Equation.2  ( AUTONUMLGL ) Obviously the numerical values can vary between (2(, as any of the terms in the difference can vary between ((. Assuming that the phase errors of two subsequent phase samples are independent random variables with identical and symmetric distributions the resulting distribution density of the phase difference is the correlation product of the individual phase distribution densities. Thus we have:  EMBED Equation.2  ( AUTONUMLGL ) Later on we will need the distribution function rather than the density. This function is simply obtained by:  EMBED Equation.2  ( AUTONUMLGL ) Conceptually equation 7 can be solved since the individual terms are known from /Mid1, LoKr1/. The phase error distribution density, given there, is:  EMBED Equation.2  ( AUTONUMLGL ) where  EMBED Equation.2  is the coherence at point k. The corresponding distribution function which we also need in equation 7, obtained from /Mid1/ is given by:  EMBED Equation.2  ( AUTONUMLGL ) As indicated earlier the phase difference  EMBED Equation.2 given in equation 5 and showing values between (2( is now mapped into the base interval between ((. The functional mapping is:  EMBED Equation.2  ( AUTONUMLGL ) The distribution density of the ((-mapped phase error difference  EMBED Equation.2 can be obtained by letting:  EMBED Equation.2  ( AUTONUMLGL ) Using the correct functional mapping in each of the intervals and exploiting the fact the conditional probability density of a functionally mapped variable only consists of a Dirac impulse we have:  EMBED Equation.2  ( AUTONUMLGL ) Now we return to the sum in equation 4 and introduce the non-mapped discrete difference by:  EMBED Equation.2  ( AUTONUMLGL ) Further introducing the conditional density of  EMBED Equation.2  conditioned on the fact that the true phase derivative takes on the value  EMBED Equation.2  we may write:  EMBED Equation.2  ( AUTONUMLGL ) In the last equality we have used the fact that the ((-mapped phase error difference  EMBED Equation.2  is independent of the true phase derivative  EMBED Equation.2 . Substituting equation 12 into 14 we obtain:  EMBED Equation.2  ( AUTONUMLGL ) This is the conditional distribution density of  EMBED Equation.2  conditioned on the fact that the true phase derivative takes on the value  EMBED Equation.2 . This variable is ((-mapped again to yield  EMBED Equation.2  (Equ. 4). Now utilizing the same arguments and reasoning as before, we get the final result for the conditional density of the mapped phase derivative, conditioned on the fact that the true phase derivative takes on the value  EMBED Equation.2 :  EMBED Equation.2  ( AUTONUMLGL ) where the short hand expression EMBED Equation.2  has been utilized for convenience. This expression is the 2(-cutout of the sum of three shifted replicas of the distribution density  EMBED Equation.2  (cf. Equ. 15):  EMBED Equation.2  ( AUTONUMLGL ) Figure 1 demonstrates the generation of  EMBED Equation.2  for an arbitrary density  EMBED Equation.2 :  EMBED Designer Figure 1: The generation of  EMBED Equation.2  as a cutout of three superimposed densities To simplify the further derivation we introduce the bias error of the mapped derivative:  EMBED Equation.2  ( AUTONUMLGL ) If we now evaluate the conditional density of this error conditioned on  EMBED Equation.2 , we can evaluate any stochastic measure of the bias conditioned on any value of the derivative, which we seek, e.g. the conditional mean of the bias. From probability theory we know that:  EMBED Equation.2  ( AUTONUMLGL ) Substituting the identity of equ. 19 into equation 16 we obtain the wanted conditional density:  EMBED Equation.2  ( AUTONUMLGL ) with  EMBED Equation.2  given in equation 17. Figure 2 demonstrates the meaning of equation 20 graphically:  EMBED Designer Figure 2: The generation of the conditional error distribution With the help of figure 2 the conditional expectation of the bias error is readily calculated:  EMBED Equation.2  ( AUTONUMLGL ) 1. As our first case we will consider the interval  EMBED Equation.2 . For this case we can subdivide the integral into the following two parts:  EMBED Equation.2  ( AUTONUMLGL ) 2. The second case is given by:  EMBED Equation.2 . Here the following sequence of operations is valid:  EMBED Equation.2  ( AUTONUMLGL ) Since  EMBED Equation.2  is an even density function symmetric around zero, the corresponding distribution  EMBED Equation.2  will show the following symmetry:  EMBED Equation.2  ( AUTONUMLGL ) From inspecting equations 22 and 23, respectively, we conclude that the conditional mean of the bias error is an odd function with respect to the nominal value  EMBED Equation.2 :  EMBED Equation.2  EMBED Equation.2  ( AUTONUMLGL ) 1.2 Preliminary Observations General observations: Summarizing equations 22-25 we note that: The bias error and true value have opposite signs The bias is an odd function with respect to the true value Thus the estimate of the phase derivative computed from modulo-mapped phases will always be biased towards lower absolute values. The phase slope is always underestimated. Limiting cases: a) Maximum Phase Slope Inspecting equations 22 and 23 we note that the bias error clearly depends on the value of the true phase slope. Knowing that:  EMBED Equation.2  ( AUTONUMLGL ) we conclude from equations 23 and 25 that:  EMBED Equation.2  ( AUTONUMLGL ) Thus phase slopes of  EMBED Equation.2 =(( will always - even for good coherence - be estimated with the maximum possible bias error of -sign( EMBED Equation.2 )((! b) Ideal Case: Perfect Correlation In this case we have:  EMBED Equation.2  ( AUTONUMLGL ) and using equations 22 and 23 we note that:  EMBED Equation.2  ( AUTONUMLGL ) c) Worst Case: No Correlation In the zero correlation case we get a uniform distribution density  EMBED Equation.2  and a ramp distribution:  EMBED Equation.2  ( AUTONUMLGL ) and from equations 22 and 23 we get the final result:  EMBED Equation.2   EMBED Equation.2  ( AUTONUMLGL ) so that in this case the estimated phase slope will always be zero, which is quite remarkable. 1.3 The General Case - Numerical Results From equations 22, 23 we conclude that knowing F0(() is completely sufficient for determining the bias error. If we furthermore restrict us to the case of phase slopes between ((, we can utilize equation 17 and write:  EMBED Equation.2  ( AUTONUMLGL ) where in the last equality we have only used the usual symmetry properties (cf. equ. 24). For convenience we will furtheron restrict ourselves to the case of positive slopes so that we can substitute equation 32 into equation 23 and write:  EMBED Equation.2  ( AUTONUMLGL ) The distribution density is periodic with respect to 4(. This means that we can expand it in a Fourier series:  EMBED Equation.2  ( AUTONUMLGL ) Substituting equation 34 into 33 we readily obtain:  EMBED Equation.2  ( AUTONUMLGL ) where the Fourier coefficients are given by:  EMBED Equation.2  ( AUTONUMLGL ) Since we do not want to solve the integral analytically we approximately calculate dm by FFT-techniques by:  EMBED Equation.2  ( AUTONUMLGL )  EMBED Equation.2  is the sampled continuous density  EMBED Equation.2  where:  EMBED Equation.2  ( AUTONUMLGL ) The continuous density  EMBED Equation.2  is, as indicated by equation 6, the continuous correlation:  EMBED Equation.2  ( AUTONUMLGL ) The discrete equivalent employing the sampled versions of the individual densities is given by:  EMBED Equation.2  ( AUTONUMLGL ) Realizing this discrete convolution as a cyclic convolution we carry over to FFT-techniques by writing:  EMBED Equation.2  ( AUTONUMLGL ) The result of equation 41 can be easily obtained in the frequency domain by letting:  EMBED Equation.2  ( AUTONUMLGL ) where the Fourier transforms are calulated by::  EMBED Equation.2  ( AUTONUMLGL ) Then the rule for approximately evaluating the bias is:  EMBED Equation.2  ( AUTONUMLGL ) Finally we obtain the solution for negative phase slopes by (equation 25):  EMBED Equation.2  ( AUTONUMLGL ) Equations 44 and 45 provide the final result and form the basic framework for evaluating the bias error depending on the phase slope itself as well as on the form of the densities. These densities depend on the degree of coherence or on the SNR of the interferogram (the quality of the fringes). In the following we will give some quantitative results. We will assume identical distribution densities for two successive points. Figure 3 shows the wrapped phase error density for different coherence values.  EMBED MgxDesigner  EMBED MgxDesigner Figure 3: Distribution Density of the Phase ErrorFigure 4: Distribution Density of the unwrapped Phase Slope ErrorFigure 4 shows the distribution density of the unwrapped phase slope error for different degrees of coherence. Figure 5 shows the outcoming bias over the true phase slope evaluated for different degrees of coherence.  EMBED MgxDesigner It is completely obvious that the maximum allowed phase slope that may be estimated with negligible bias strongly depends on the coherence. If the coherence is one, there is no phase slope bias as long as the phase slope is less than (. The other extreme is a coherence of 0.1. In this case the slope bias is considerable even for small slopes.Figure 5: Bias Error over Phase Slope2.0 Approaches to solve or circumvent the problem The following chapter will give a short overview about how to avoid or circumvent the problem of biased phase derivative estimates. Do not apply any filtering or averaging techniques to the phase slope! Any filter operation to remove the stochastic influences from the phase derivative will produce an estimate which is nearer to the conditional mean, which is not identical with the true phase slope. On the other hand the simple branch/cut methods which do not apply any filtering to the phase slope estimates provide noisy but unbiased phase estimates. If any filtering is to be applied it should be applied to complex data rather than to the phases or phase slopes. Keep the phase slopes as small as possible by successive flattening techniques! Since the bias of the slope estimation clearly depends on the slope itself, one method to keep the bias small might be to keep the phase slope small. This can be achieved by successive flattening. The biased phase estimates obtained in the first run are utilized to demodulate the complex interferogram. Then the residual phase slope is estimated again, this time with a smaller bias. The unambiguous phase is generated and used again to demodulate the interferogram. The whole loop is repeated until the bias has been reduced satisfactorily. Such a procedure (as reported in /For1 / clearly yields asymptotically unbiased phase estimates. Correct the systematic error by subtracting the phase slope bias! With the results given in equations 44 and 45 it should be possible to further utilize the phase slope estimator based on finite differences of wrapped phases and to eliminate the bias by simply subtracting it. This method would be advantageously applicable to homogeneous scenes with a sufficiently smooth coherence distribution, since then the densities would not have to be calculated and Fourier transformed for any individual pixel. In this case the additional computational burden would be moderate. The bias estimation of equations 44 and 45 would be the key to maintain Linear Least Squares phase unwrapping approaches. Do not calculate phase slopes by finite differences from modulo mapped phases! Clearly the best solution to a problem is an approach which prevents the problem from arising. This can be achieved by applying unbiased phase slope estimators. All these estimators share the common property that they operate on complex data rather than on the phases. Either they use real and imaginary part of the interferogram and exploit the argument of a complex correlation kernel (known from Madsens Correlation Doppler Estimator, as proposed by /Bam1/) or they operate in the power spectral density domain (such as the Local Spectral Mode Estimator, proposed by /KrLo2, LoKr2/). Another approach would be to use a nonlinear estimator in form of an Extended Kalman Filter which does not explicitly differentiate any mapped phases (as proposed in /LoKr1, KrLo1/). Recently a combination of local slope estimation and Kalman filtering techniques has been proposed in /KrLo2, LoKr2/. This combination seems to be the most powerful approach to phase unwrapping, yielding unbiased and nearly perfectly noisefree unwrapped phases down to coherence values of 0.3 without any prefiltering! Conclusions The paper has presented the analysis, the derivation and evaluation of the estimation bias if the phase slope is determined form modulo-2( mapped phases. A finite series representation for the phase slope bias has been given, where the coefficients can be easily determined from the FFT spectrum of the distribution density of the wrapped phase error. This kind of distribution density is known for a lot of special cases even if multi-look prefiltering is applied. Thus the technique is widely applicable. The numerical results calculated with the approach are in perfect agreement with the expectation. Finally some hints to circumvent or solve the problem were presented. Acknowledgment The work reported in the paper is an integral part of our activities on phase unwrapping in cooperation with DLR and has been funded by the DARA (project Rapid, grant number 50 EE 9431) which is greatly appreciated. References Middleton, D., 1987: Introduction to Statistical Communication Theory, Peninsula Publishing, Los Altos, pp. 396-410. ( Mid1 ) Loffeld, O., Krmer, R., 1994 Phase Unwrapping for SAR Interferometry, Proc. IGARSS94, Pasadena,, pp. 2282-2284, (LoKr1) Krmer, R., Loffeld, O., 1996 Phase Unwrapping for SAR Interferometry with Kalman Filters , Proc. EUSAR96, Knigswinter, pp. 199-202, (KrLo1) Krmer, R., Loffeld, O., 1996 Local Slope Estimation in cooperation with Kalman Filtering Techniques, DLR Workshop on Phase Unwrapping, (KrLo2) Loffeld, O., Krmer, R., 1996 Local Slope Estimation and Kalman Filtering, PIERS96 Symposium, Innsbruck, 1996 (LoKr2). Bamler, R., Davidson, G., 1996 The mystery of lost fringes in 2D-Least-Squares Phase Unwrapping, Piers96, Symposium, Innsbruck, 1996. (Bam1) Fornaro, G., Franceschetti, G. Lanari, R., Rossi,D., Tesauro, M., 1996 Phase Unwrapping by using the Finite Element Method (FEM), Piers96, Symposium, Innsbruck, 1996. (For1)   PRIVATE href="http://www.geo.unizh.ch/rsl/fringe96/program.html"MACROBUTTON HtmlResAnchor (Conference Program)  PRIVATE src="http://www.geo.unizh.ch/rsl/fringe96/icons/space.gif" width=5 height=1 alt="" MACROBUTTON HtmlResImg INCLUDEPICTURE "C:/WIN95/TEMP/wia8b0/space.bmp" \* MERGEFORMAT    PRIVATE href="http://www.geo.unizh.ch/rsl/fringe96/participants/"MACROBUTTON HtmlResAnchor (Participants)  PRIVATE src="http://www.geo.unizh.ch/rsl/fringe96/icons/space.gif" width=5 height=1 alt="" MACROBUTTON HtmlResImg INCLUDEPICTURE "C:/WIN95/TEMP/wia8b0/space.bmp" \* MERGEFORMAT    PRIVATE href="http://www.geo.unizh.ch/rsl/fringe96/papers/"MACROBUTTON HtmlResAnchor (Abstracts and Papers)  PRIVATE src="http://www.geo.unizh.ch/rsl/fringe96/icons/space.gif" width=100 height=1 alt="" MACROBUTTON HtmlResImg INCLUDEPICTURE "C:/WIN95/TEMP/wia8b0/space.bmp" \* MERGEFORMAT    PRIVATE href="http://www.geo.unizh.ch/rsl/fringe96/contacts.html"MACROBUTTON HtmlResAnchor (Contacts) /= h        .........)()()d :*    O C OO(O!!!555NNNbbbh Hh Hrh=feth":ptw//.wwoegnu.hzihc.sr/f/lnir9egc/6tnotcah.slmt"                                                                                                                          e  :    / C //(/@  !!!"""###$$$%%%&&&'''((()))***+++,,,---...///000111222333444555666777888999:::;;;<<<===>>>???@@@AAABBBCCCDDDEEEFFFGGGHHHIIIJJJKKKLLLMMMNNNOOOPPPQQQRRRSSSTTTUUUVVVWWWXXXYYYZZZ[[[\\\]]]^^^___```aaabbbcccdddeeefffggghhhiiijjjkkklllmmmnnnooopppqqqrrrssstttuuuvvvwwwxxxyyyzzz{{{|||}}}~~~D: Bk8"MZo^">.J">J">oA3J>DU0J>`J>*J>3-J>K*,lJ>ZHdJ>:9"J>^0dJ>4l5J> 3J>3J>J>JWv>s@x=qg: :V  nJ  J JC JJ(JPRRRZZZcc!c)k)k1k9s9sBsJ{J{RZcks{ 2 2  2 2 2 2 2 2 2  2 2 2 2 2 2 2  2    2      2        2      2     2   2   2 2 2  2 2  4:11  H$   $C (210543876;:9>=<A@?DCBGFEJIHMLKPONSRQVUTYXW\[Z_^]ba`edchgfkjinmlqpotsrwvuzyx}|{~UUVVVWWWXXXYYYZZZ[[[\\\]]]^^^___```aaabbbcccdddeeefffggghhhiiijjjkkklllmmmnnnooopppqqqrrrssstttuuuvvvwwwxxxyyyzzz{{{|||}}}~~~ Jro :    N C NN(N!!!111FFFZZZkkksss{{{BIL\RUPENNUT0ESUSR0GAHREG0enectret0csluhre#2noKpezS tuhcrelod.c<J@<J@<J@i:ע+00@\XE\LACUBIR\\CXEILARUBEN\UTPN0SUSRE0AHEGGR0ncteeter0suhcrelJ,`: i00+@E\\ACXBILRU\XE\LACUBIN\RTPENU0URESS0HGGARE0tentecre0hcselurYKKKKKKKKKKKKKKKK       K     K    K    K     K       K      K K  KKKKnc:11  H$   $C (210543876;:9>=<A@?DCBGFEJIHMLKPONSRQVUTYXW\[Z_^]ba`edchgfkjinmlqpotsrwvuzyx}|{~UUVVVWWWXXXYYYZZZ[[[\\\]]]^^^___```aaabbbcccdddeeefffggghhhiiijjjkkklllmmmnnnooopppqqqrrrssstttuuuvvvwwwxxxyyyzzz{{{|||}}}~~~Root Entry=ΦLΦ[NWordDocumentnObjectPool$`p=Φ`p=ΦSummaryInformation(  !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOvYSTUVG]\^_`abcdefghijklmnopqxyz{|}~de|^_wx  $%<>@AB   ?@WY[]^nouDC^b^b uD^b uDraUb c uDeaPC uDCuDRNOgh  $%defhijlm2 3 K L 6 7 N P Q ` a c d { | M N P Q h i B C P uDab uDaC uDCuD uD|ab uD^b^buDC^bC^bNC E F ] ^ u w x + , . / F G ^ ` a n IJ 68ce  HI\]^_Ľ嵬uDL5a eKuDL5a vK uDM5eK uDM5vKJjuDO5a eKuDO5a vK uDa a hVJpUUuDPC uDCD   ijklmOPQdefgijvwy0123O P Q d e f g i j v w y G!H![!\!]!^!!!""% uDG5eK uDG5vKuDH5a eKuDH5a vK uDJ5eK uDJ5vKuDuDK5a eKuDK5a vKJhJ]Jp uDa a B%%%%%%%%%*%+%-%]%^%_%`%%%%&&&&&&&&&&&&U'V'W'j'k'l'm'o'p'|'}''''(((*(+(,(-(/(0(<(=(?(E(F(Y(Z(Ǿ uDA5vKuDuDB5a eKuDB5a vKcuDD5a eKuDD5a vKuDE5a eKuDE5avKJpJuDF5a eKuDF5a vK uDa a ?@ABCDEFGHIJKLMNOsY]\^_`abcdefghijklmnopqxtuvwryz{|}~DocumentSummaryInformation8_904720975>F`p=Φ`p=Φ_904720973:F z=Φ z=Φ_9047209724F z=Φ@=Φ  !"$&'()*+,-./0123456789:<@ABCDEFGHIJKLMNOPQSUYZ[\]^`bcdefghijklmnopqrstuvwxyz{|}~՜.+,0HPlt|  CAMPUS Systems AG8) Φ_904720948F G>Φ@>Φ_904720947 F@>Φ>Φ_904720945F>Φ >Φ_904720944F >Φ!>Φ_904720943F!>Φz)>Φ_904720942"Fz)>Φ2>Φ_904720941F2>Φ2>Φ_904720939#!F2>ΦB:>Φ_904720938FB:>ΦjC>Φ_9047209374FjC>Φ K>Φ_904720936F K>Φ3T>Φ_904720935'%F3T>Φ[>Φ_904720933}F[>Φ`d>Φ_904720932*&w1F`d>Φl>Φ_904720931qFl>Φ@u>Φ_904720930+)kF@u>Φ@u>Φ_904720928eF@u>Φ`f}>Φ_9047209270(_F`f}>Φ >Φ_904720926YF >Φ@/>Φ_904720925/-SF@/>ΦW>Φ_904720923M1FW>Φ >Φ_9047209222.GF >ΦVxAΦ_904720921AFVxAΦAΦ_90472092031;FAΦAΦ_9047209195FAΦAΦ_904720917D,/FAΦAΦ_904720916)FAΦAΦ_90472091575#FAΦAΦ_904720914FAΦRAΦ_904720913:6FRAΦ`zAΦ_904720911F`zAΦ`zAΦ_904720910;9 F`zAΦAΦ_904720909FAΦ@CAΦ_904720908@8F`AΦ`AΦ_904720907F`AΦ AΦ_904720905?=F AΦ@AΦ_904720904F@AΦAΦ_904720903B>FAΦ vAΦ_904720902F vAΦAΦ_904720901CAFAΦ?BΦ_904720899F?BΦfBΦ_904720898L<FfBΦBΦ_904720897FBΦ/ BΦ_904720896GEF/ BΦ'BΦ_904720895F'BΦ0BΦ_904720893JFF0BΦ8BΦ_904720892F8BΦ`ABΦ_904720891KIF`ABΦbIBΦ_904720889FbIBΦ@RBΦ_904720888PHF@RBΦ`+ZBΦ_904720887F`+ZBΦ ScBΦ_904720886OMF ScBΦ@jBΦ_904720885F@jBΦtBΦ_904720883TNFtBΦ {BΦ_904720882F {BΦBΦ_904720881SQ~FBΦBΦ_904720880xFBΦDΦ_904720878VRrFDΦ/EΦ_904720877lF/EΦ@W EΦ_904720876WUf]F@W EΦ`EΦ_904720875X`]F EΦ@"EΦ_904720873Z]F@"EΦ+EΦOle PIC Y\LMETA Xq J Ja:11  H$   $C (210543876;:9>=<A@?DCBGFEJIHMLKPONSRQVUTYXW\[Z_^]ba`edchgfkjinmlqpotsrwvuzyx}|{~UUVVVWWWXXXYYYZZZ[[[\\\]]]^^^___```aaabbbcccdddeeefffggghhhiiijjjkkklllmmmnnnooopppqqqrrrssstttuuuvvvwwwxxxyyyzzz{{{|||}}}~~~ J:11  H$   $C (210543876;:9>=<A@?DCBGFEJIHMLKPONSRQVUTYXW\[Z_^]ba`edchgfkjinmlqpotsrwvuzyx}|{~UUVVVWWWXXXYYYZZZ[[[\\\]]]^^^___```aaabbbcccdddeeefffggghhhiiijjjkkklllmmmnnnooopppqqqrrrssstttuuuvvvwwwxxxyyyzzz{{{|||}}}~~~CompObj[^ [ObjInfo Ole10Native]$Ole PIC _bLMETA jCompObjad[ObjInfoOle10Nativec$Ole PIC ehLMETA YiCompObjgj[ObjInfoOle10Nativei $Ole =PIC kn;LMETA %HCompObjmp#_ObjInfoEquation Native oOle VPIC qtTLMETA CompObjsvR_ObjInfo>Equation Native u?Ole PIC wzLMETA ah CompObjy|__ObjInfoWEquation Native {XOle PIC }LMETA CompObj_ObjInfoEquation Native Ole PIC LMETA CompObj_ObjInfoEquation Native Ole PIC LMETA CompObj_ObjInfoEquation Native Ole PIC LMETA HCompObj_ObjInfoEquation Native Ole (PIC &LMETA CompObj_ObjInfoEquation Native \Ole LPIC JLMETA 2CompObj0_ObjInfo)Equation Native *\Ole `PIC ^LMETA RCompObjP_ObjInfoMEquation Native N\Ole sPIC qLMETA d(ObjInfoaEquation Native b\Ole PIC LMETA ( CompObj~_ObjInfotEquation Native u<Ole PIC LMETA CompObj_ObjInfoEquation Native Ole PIC LMETA CompObj_ObjInfoEquation Native Ole PIC LMETA HCompObj_ObjInfoEquation Native Ole VPIC TLMETA (CompObj_ObjInfoEquation Native  Ole oPIC mLMETA CompObjk_ObjInfoWEquation Native XOle PIC LMETA vCompObjt_ObjInfopEquation Native qOle PIC LMETA CompObj_ObjInfoEquation Native <Ole PIC LMETA hCompObj_ObjInfoEquation Native Ole PIC LMETA CompObj_ObjInfoEquation Native Ole PIC LMETA HCompObj_ObjInfoEquation Native Ole PIC LMETA CompObj_ObjInfoEquation Native Ole PIC LMETA hObjInfoEquation Native \Ole  PIC LMETA hObjInfoEquation Native \Ole HPIC FLMETA )CompObj'_ObjInfo!Equation Native "Ole `PIC ^LMETA OCompObj M_ObjInfoIEquation Native JOle PIC   ~LMETA hHCompObj f_ObjInfoaEquation Native bOle PIC LMETA CompObj_ObjInfoEquation Native <Ole PIC LMETA hCompObj_ObjInfoEquation Native <Ole PIC LMETA CompObj!_ObjInfoEquation Native  |Ole PIC "%LMETA CompObj$'_ObjInfoEquation Native &\Ole PIC (+LMETA CompObj*-_ObjInfoEquation Native ,|Ole $PIC .1"LMETA CompObj03_ObjInfoEquation Native 2|Ole 5PIC 473LMETA *(CompObj69(_ObjInfo%Equation Native 8&\Ole KPIC :=ILMETA =CompObj<?G_ObjInfo6Equation Native >7Ole \PIC @CZLMETA Q(CompObjBEO_ObjInfoLEquation Native DM\Ole PIC FILMETA eCompObjHKc_ObjInfo]Equation Native J^Ole PIC LOLMETA _CompObjNQUObjInfoOle10NativePN$ Ole PIC RULMETA CompObjTW_ObjInfoEquation Native V\Ole PIC X[LMETA CompObjZ]_ObjInfoEquation Native \<Ole PIC ^aLMETA tCompObj`c_ObjInfoEquation Native bOle PIC dgLMETA CompObjfi_ObjInfoEquation Native h\Ole PIC jmLMETA CompObjlo_ObjInfoEquation Native n|Ole PIC psLMETA CompObjru_ObjInfoEquation Native t\Ole VPIC vyTLMETA CompObjx{UObjInfoOle10Nativez$Ole jPIC |hLMETA ZHObjInfo~WEquation Native X\Ole {PIC yLMETA pCompObjn_ObjInfokEquation Native l\Ole PIC LMETA (CompObj_ObjInfo|Equation Native }<Ole PIC LMETA HObjInfoEquation Native \Ole PIC LMETA CompObj_ObjInfoEquation Native \Ole PIC LMETA HCompObj_ObjInfoEquation Native Ole !PIC LMETA hCompObj_ObjInfoEquation Native <Ole 2PIC 0LMETA 'CompObj%_ObjInfo"Equation Native #\Ole ?PIC =LMETA 7hCompObj5_ObjInfo3Equation Native 4<Ole MPIC KLMETA BObjInfo@Equation Native A<Ole PIC LMETA X CompObjV_ObjInfoNEquation Native OOle PIC LMETA CompObj_ObjInfoEquation Native \Ole PIC LMETA hObjInfoEquation Native \Ole PIC LMETA CompObj_ObjInfoEquation Native <Ole PIC LMETA hCompObj_ObjInfoEquation Native <Ole PIC LMETA ObjInfoEquation Native <Ole PIC LMETA CompObj_ObjInfoEquation Native Ole PIC LMETA HCompObj_ObjInfoEquation Native  |Ole 6PIC 4LMETA hCompObj2_ObjInfo Equation Native !<Ole FPIC DLMETA :hObjInfo7Equation Native 8\Ole wPIC uLMETA QCompObjO_ObjInfoGEquation Native HOle PIC LMETA {hObjInfoxEquation Native y\Ole PIC LMETA  CompObj_ObjInfoEquation Native Ole PIC LMETA CompObj_ObjInfoEquation Native \Ole PIC LMETA CompObj _ObjInfoEquation Native  Ole 3PIC  1LMETA ( CompObj _ObjInfoEquation Native <Ole ]PIC [LMETA =hCompObj;_ObjInfo4Equation Native 5\Ole uPIC sLMETA dCompObjb_ObjInfo^Equation Native _Ole PIC  LMETA yhObjInfo!vEquation Native w\Ole PIC "%LMETA H CompObj$'_ObjInfoEquation Native &Ole PIC (+LMETA ObjInfo*,Equation Native |Ole PIC -0LMETA 3hCompObj/2_ObjInfoEquation Native 1Ole  PIC 36 LMETA CompObj58_ObjInfoEquation Native 7Ole  PIC 9; LMETA < ObjInfo Ole : PIC =@8 LMETA ' CompObj?B% _ObjInfo! Equation Native A"       !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~ RIFFMGX itpddrwlJa FPXRDE0@@PR''edSS$ U=```@ap@`:@8c@ b 0p1 F@!#   @#@ @ @ @ @ @@@~?|?0    ?:    /#@!< o ? 8@ < d> pp? >N0 ? p~  @8`8#0< Ap  @ 8 x hB 1|  8x< f x ` 0 p`  p ` `888 <   `@ @ 0p   `   ?0`p  @ `8? @@  `8 ɀ `@p 00`A8`<   0 (  0   8 8h 0< `8  @`  @ d `    @@ @8`< @`p 4  ` @@0@ @ 4`08< @ @ @`0 0       @  0@ ?@0   "@  `?            b@` `  A8  <     `@ @ 8   @ 8@@@ < 10   `  @  1@  0@ @@ 8< @  @ @@ @ 00@ 0 8`<p @  @@@    ` @     @@ ?   ` @  0 ?         @  @  0 A8<  g  @ @ 0@   @@#  " "8<" @" A0? @       @8@` < 0   `   @   @8`@<   @`  @ D0@  0  A  ?` @ @ 0@? @   @ @   ` A8 @@<   A@  ` @ @ 0 A 8@@< @  d     @ @` `  @8  @< D@@   @  $     C 8< @@ 4`@0   @     "`@ @ ?     ? @   @   @A8a <       a@  8 <  @ `  0 @  !   d@80D <     `B -!  D8H<  $  "   A A   iD?0  a? ҄ y d ` `A8 l     <??<?~@ BA@ BA@? @  @ @ @ B @B c??<1? H H K $D"?$8D $$ $$ $D":aYQPZRDE0@@PR H!z SR drbb{ w k BIAS.DS4#2#1996820180213 Equation.2 ࡱ; Root EntryFԪt@Equation Native <~ @t 掛  g=0.1ࡱ;  METAFILEPICT4l  .1  &@ & MathTypePSymbol- 2 `%gSymbol- 2 `q=Times- 2 `0 2 `1 2 `^.`&,MathTypeUU  g=0.1 &  "System-draduq[QPZRDE0@@PR0oPF1=\SR drbb{ w k BIAS.DS4#2#1996820180513 Equation.2 ࡱ; Root EntryFJ[@Equation Native <~  vl掛 l g=1.0ࡱ;  METAFILEPICT4nl  .1  &@ & MathTypePSymbol- 2 `%gSymbol- 2 `q=Times- 2 `1 2 `y0 2 `.`&,MathTypeUU  g=1.0 &  "System-draduq[QPZRDE0I@v@PR4YioSR drbb{ w k BIAS.DS4#2#1996820181113 Equation.2 ࡱ; Root EntryF G@Equation Native <~ @wl掛 l g=0.9m mehrerࡱ;  METAFILEPICT<<46l  .1  & & MathTypePSymbol- 2 `%gSymbol- 2 `q=Times- 2 `0 2 `9 2 `^.`&,MathTypeUU  g=0.9s Objekt &  "System-draduq[QPZRDE0@R@PR(Hq+\5SR drbb{ w k BIAS.DS4#2#1996820181323 Equation.2 ࡱ; Root EntryFc@Equation Native <~ kl掛 l g=0.8ࡱ;  METAFILEPICT4l  .1  &}`} & MathTypePSymbol- 2 `%gSymbol- 2 `q=Timeso- 2 `0 2 `8 2 `^.`&,MathTypeUU  g=0.8>? &  "System-draduq[QPZRDE0 @@PR0,eSR drbb{ w k BIAS.DS4#2#1996820181573 Equation.2 ࡱ; Root EntryF@Equation Native <~ zk掛 k g=0.7ࡱ;  METAFILEPICT<<4k  .1  &: : & MathTypePSymbol- 2 `%gSymbol- 2 `q=Times- 2 `0 2 `7 2 `^.`&,MathTypeUU  g=0.7chnen ei &  "System-draduq[QPZRDE0@@PRP@tSR drbbz v j BIAS.DS4#2#199682018333 Equation.2 ࡱ; Root EntryF@Equation Native <~  tl掛 l g=0.6ࡱ;  METAFILEPICT<<4nk  .1  &  & MathTypePSymbol- 2 `%gSymbol- 2 `q=Times- 2 `0 2 `6 2 `^.`&,MathTypeUU  g=0.6s Objekt &  "System-draduq[QPZRDE0@p@PR Hk{mSR drbb{ w  k BIAS.DS4#2#1996820183293 Equation.2 ࡱ; Root EntryF-@Equation Native <~ }l掛 l g=0.5  uࡱ;  METAFILEPICT4k  .1  &` & MathTypePSymbol- 2 `%gSymbol- 2 `q=Times- 2 `0 2 `5 2 `^.`&,MathTypeUU  g=0.5ehen, um &  "System-drad [QG{lp f B` BQ ;    "H  yj d B[ RH J;t 5 B/ B       "YH P H9 3 B* !H  x  B B# "# # "x# "d# "O# "#R# ""T""! ! {! m!> `!> S! !> !V >  >      H  F t D ^  T  G  2  "        >' ' B' 'H ''x ' B' B' y' P' ,' ' & & & & "K&D B& :&+& %& B& &D  &%X % % B% %F % %% % B% %F %%h % }% Bn% J%H A% 9%*% $% B% %H  %$| $ B$ B$ $ $ ~$ a$ 9$ #$  $ + +H + "3f3333f333ff3fffff3f3f̙3f3333f3333333333f3333333f3f33ff3f3f3f3333f3333333f3̙333333f333ff3ffffff3f33f3ff3f3f3ffff3fffffffffff3fffffff3fff̙ffff3fffff3f̙3333f33̙3ff3ffff̙f3f̙3f̙̙3f̙3f3333f333ff3fffff̙̙3̙f̙̙̙3f̙3f3f3333f333ff3fffff3f3f̙3f h*  8 &0&/MGXI: created by Micrografx, Inc. Mgxgre v1.00    &0C (` $D" $$ $$?$8D $D" K H H?c??<1 @B  @ B @ @ @ @?@ BA@ BA?<?~?     < l`A8`   d y? ҄ a0 ? iD A  A  "    $8H< D  -! `B    80D < d@    ! @ 0 `  @ 8 <   a@      A8a < @ @    ? @   ?  @  "`@     @   4`@0 @@8< C      $ @    D@@8  @< @ `  @` @    d  @ 8@@< A  0 @ @  ` A@  A8 @@<`   @   @ ? @  0@ @ @?`   A 0  D0@  @  @`  8`@< @  @    `   08@` < @      ? @" A0" @"8<" #    @@ 0@  @ @g   A8< 0   @    @    ?    0   ` @ ?  @@     ` @     @@@p @8`<  0 00@ @@ @ @  @8< @@  0@ 1@  @   `   108@@@ < @    8 @ `@    A8  < `    b@`       ?     `  "@ ?@0 0@   @      @`0 0 @ @8< 4`0 @ @@0@ `  4 @`p8`< @@ @   d `   @ @`  `88h 0<  8   0 (  0  `A8`< 00 `@p ɀ  `8? @@ `8  @?0`p     `  0p @ `@  8 < `88 p `  p` 0 ` f x8x<   1| hB x @ 8  Ap8#0< @8`  p~ ? >N0 pp? d>8@ <  ? o!<#@/?:    0    ?~?|@@@ @@ @ @ #@ @!#   @c@ b 0p1 Fp@`:@8a--  "System- rj .1 '5& &@ & MathTypePSymbol- 2 `%gSymbol- 2 `q=TimesC- 2 `0 2 `1 2 `^.`&,MathTypeUU  g=0.1 &  "System-'  -- - w^ .1 5& 2p&@ & MathTypePSymbol- 2 `%gSymbol- 2 `q=Times- 2 `1 2 `y0 2 `.`&,MathTypeUU  g=1.0 &  "System-'  -- - / .1 'L& U& & MathTypePSymbol- 2 `%gSymbol- 2 `q=Times- 2 `0 2 `9 2 `^.`&,MathTypeUU  g=0.9s Objekt &  "System-'  -- - #  .1 A& P&}`} & MathTypePSymbol- 2 `%gSymbol- 2 `q=Times- 2 `0 2 `8 2 `^.`&,MathTypeUU  g=0.8>? &  "System-'  -- - u3# .1 'L& &: : & MathTypePSymbol- 2 `%gSymbol- 2 `q=Times- 2 `0 2 `7 2 `^.`&,MathTypeUU  g=0.7chnen ei &  "System-'  -- - zR .1 'L& &  & MathTypePSymbol- 2 `%gSymbol-  2 `q=Times- 2 `0 2 `6 2 `^.`&,MathTypeUU  g=0.6s Objekt &  "System- '  -- - j .1 'A& e&` & MathTypePSymbol- 2 `%gSymbol-  2 `q=Times- 2 `0 2 `5 2 `^.`&,MathTypeUU  g=0.5ehen, um &  "System- '    ]FDesigner 4.1 Zeichnung MgxDesigner MgxDesignerLh*   RIFF(yMGX itpddrwlRa FPXRDE0@^@PR8n`SS$ U=```@  1 ~0!A@a@@@D>!p O@ \| 81`p,Da! B0@!b B@ `# `D0DA! B`@"a"A@ D@! B@@$A! D>! B@D@'A" ! B@D@DAB ! B@D@BA"@ !@@BA" A ~?? >| `? @@@@  @@@@ @ @8 @<D   8<H ( $ D ᇃ8<   8 <   8@< @ @  8<   8< @ @@ @ @8@< @ @ 8 <    ?  H ( (8 <$ D ៃ 8<    8<   8<   8<    8<   8<    8 <   8<   ?    8@< @ @ @8@< @ @ @@ @8 <     8<  @ @8@<        8 <    8<     @ @8@< @ @ @8H<   ?    @H @(  (8 <$  D  ៃ  8<       8@<  ! !8A!< b @ @8@< @ ` ` 08 <   8<      80< 0 ? @88<  @ D?$D D H D  T0  @8D8a`< 8 0` 08(<  @ @B   088<  p 8@d @@A$8A 9@< @` p  @xh8 `<(8 x@Bpt| <( '|pFD 8x@q? <D#>1p8^@@>?? Ꮓ?@>@ @!@@ @ @ L @ r@ a ?A@ !@ !>? H H K $D"?$8D $$ $$ $D"BaYQPZRDE8U|B?@R@PR@ M@SR drbbEmbedded Object #1#199689145366 Equation.2 ࡱ; Root EntryF6Equation Native k 掛  g=0.9g=0.8g=0.7g=0.6g=0.5g=0.4g=0.3g=0.2g=0.1ࡱ; ࡱ;  METAFILEPICT<h</  k .1   & & MathTypeSymbol- 2 Q%g 2 %g 2 %g 2 %g 2 ` %g 2 %g 2 %g 2 )%g 2 l%gSymbol- 2 Qk= 2 k= 2 k= 2 k= 2 ` k= 2 k= 2 k= 2 )k= 2 lk=Times- 2 Q0 2 Q9 2 0 2 8 2 0 2 7 2 0 2 6 2 ` 0 2 ` 5 2 0 2 0 2 3 2 )0 2 )2 2 l0 2 l1 2 QR.` 2 R.` 2 R.` 2 R.` 2 ` R.` 2 R.4` 2 R.` 2 )R.` 2 lR.`k&MathTypeUU g=0.9g=0.8g=0.7g=0.6g=0.5g=0.4g=0.3g=0.2g=0.1 &  "System-drad[QP\RDEPRL$S $ddddg ptqLuq]QP\RDEPRL*hVS $ddddg ptqL*hUuq]QP\RDEPRL@hoFS $ddddg ptqL@hFuq]QP\RDEPR Hh2S $ddddg ptq Hh2uq]QP\RDEPRL B^`S $ddddg ptqL Bx]uq]QP\RDEPRLUfS $ddddg ptqLU`auq]QP\RDEPRL$`ȩmS $ddddg ptqLmȩHeuq]QP\RDEPR HU~xcS $ddddg ptq Hxcuq]QP\RDEPRL`;S $ddddg ptqL`Xuq]QG2g=0.1ࡱ; ࡱ;  METAFILEPICT<h</  k .1   & & MathTypeSymbol- 2 Q%g 2 %g 2 %g 2 %g 2 ` %g 2 %g 2 %g 2 )%g 2 l%gSymbol- 2 Qk= 2 k= 2 k= 2 k= 2 ` k= 2 k= 2 k= 2 )k= 2 lk=Times- 2 Q0 2 Q9 2 0 2 8 2 0 2 7 2 0 2 6 2 ` 0 2 ` 5 2 0 2 0 2 3 2 )0 2 )2 2 l0 2 l1 2 QR.` 2 R.` 2 R.` 2 R.` 2 ` R.` 2 R.4` 2 R.` 2 )R.` 2 lR.`k&MathTypeUU g=0.9g=0.8g=0.7g=0.6g=0.5g=0.4g=0.3g=0.2g=0.1 &  "System-  v     !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~h*l 4&0&/MGXI: created by Micrografx, Inc. Mgxgre v1.00    &0C (` $D" $$ $$?$8D $D" K H H?>@ !@ ! ?A@ a @ r@ L@@ @ @ @!@>Ꮓ?D#>1p8^@@>?? D 8x@q? <( '|pF(8 x@Bpt| <( >`<( x0`0@@D @ `D8@ 0 `<  0   p < 0p@p 8 @`p@A 80p8  < `ax p@0 080`00< `    ` ` a a00!8  `8`` 0< 1A8  @! @080` @< `p 100B ` F  8 @ !<  0F` Aa q@2L   08@ < 0@# 20@ !@p80@ !< ` ` @^ ;d 0 ˀ ?xA@   B0  @@ @88 @<  @   2/8_v<  r /r 82<   ! <8 880< D CpX t@  <<8 |!@@BA" A ! B@D@BA"@ ! B@D@DAB D>! B@D@'A" D@! B@@$A! DA! B`@"a"A@ Da! B0@!b B@ `# `D0D>!p O@ \| 81`p,@@@!A@a 1 ~0 --  "System- ^ .1 {L& & & MathTypeSymbol- 2 Q%g 2 %g 2 %g 2 %g 2 ` %g 2 %g 2 %g 2 )%g 2 l%gSymbol- 2 Qk= 2 k= 2 k= 2 k= 2 ` k= 2 k= 2 k= 2 )k= 2 lk=Timesg- 2 Q0 2 Q9 2 0 2 8 2 0 2 7 2 0 2 6 2 ` 0 2 ` 5 2 0 2 0 2 3 2 )0 2 )2 2 l0 2 l1 2 QR.` 2 R.` 2 R.` 2 R.` 2 ` R.` 2 R.4` 2 R.` 2 )R.` 2 lR.`k&MathTypeUU g=0.9g=0.8g=0.7g=0.6g=0.5g=0.4g=0.3g=0.2g=0.1 &  "System-'  --%vv-- $cvc--%-- $--%K-- $--%-- $--%+-- $+4--%7-- $ 7E --%ZC-- $3CY3--%-- $--%?-- $?B--A8  ]FDesigner 4.1 Zeichnung MgxDesigner MgxDesignerLh*   _T|'S+S Ee(k)/d RIFFzMGX itpddrwl^a FPXRDE0@^@PR8n`SS$ U=`,``tt`" " "" F" %P$P$  ??   @0@ @     ???]  A A @A A @A   @  @?!@ <x > ?? @?  |?pp@|  `x8?<  # ?? ? G 0?  @?D @~$ xx` Ā "8@  8`0@ 8p?   G@~8`  < @p ?0 00!@"! B`pB  ``@ B    8@  @@0?@@@@! 8@0B     08?@@ D< `@@`2 xa p! l b ?H@ G`O    @ # @?N @x@ (  H  0  ?B / @ !` B " @!@,D @pO x &    0  ( 8 8< r  `*     @@ (  8  P 8` P A@0 A  !8<     @ @D    H @X H0 $  8`< ` 0 0 `@ @ @ ( @ @ `@ @@ 0?" &@ @D @L            8<    @@@ @@@    `@ P@ P@ 0@ 8 <            8"< " " " @B DD       ? `             8@< @  @               8< @ @@ @@ @ @ @ @ @ @ @   8 <     D       ? `           8<               8<       @ @ @@ @@ @@ @@8@@<                 @ @ @( A  A A @A A @A ( A  A A @A A @A ?@0 ~`@?  " @`    @ @"  D   @ @"  @D   @ @" ?<@D |? @!@"  B@@  @ @"@1#0C 00 @!@ A )b(8801 | p&'<&,@ @  @  `@`0   NaYQPZRDE8'挦?@@PR@MSR drbbPH_DENS.DS4#1#19968101229427 Equation.2ࡱ; Root EntryF@Ole PIC LMETA  ࡱ; L< ࡱ; <  9 .1  @& & MathType`Symbol- 2 F%g 2 %g 2 %g 2 %g 2 U %g 2 %g 2 %g 2 %g 2 a%gSymbol- 2 Fk= 2 k= 2 k= 2 k= 2 U k= 2 k= 2 k= 2 k= 2 ak=Times- 2 F0 2 F9 2 0 2 8 2 0 2 7 2 0 2 6 2 U 0 2 U 5 2 0 2 0 2 3 2 0 2 2 2 a0 2 a1 2 FR.` 2 R.` 2 R.` 2 R.` 2 U R.` 2 R.4` 2 R.` 2 R.` 2 aR.` &  "System-g=0.1ࡱ FMicrosoft Formel-Editor 2.0 DS Equation Equation.2ࡱ; CompObj_ObjInfoEquation Native ࡱ; ddx d g=0.9g=0.8g=0.7g=0.6g=0.5g=0.4g=0.3g=0.2g=0.1ࡱ;  METAFILEPICT<y<~ 9 .1  @& & MathType`Symbol- 2 F%g 2 %g 2 %g 2 %g 2 U %g 2 %g 2 %g 2 %g 2 a%gSymbol- 2 Fk= 2 k= 2 k= 2 k= 2 U k= 2 k= 2 k= 2 k= 2 ak=Times- 2 F0 2 F9 2 0 2 8 2 0 2 7 2 0 2 6 2 U 0 2 U 5 2 0 2 0 2 3 2 0 2 2 2 a0 2 a1 2 FR.` 2 R.` 2 R.` 2 R.` 2 U R.` 2 R.4` 2 R.` 2 R.` 2 aR.` &  "System-g=0.1drad[QP\RDEPR HXFS $ddddg ptq HXFuq]QP\RDEPR H` S $ddddg ptq H`p uq]QP\RDEPR H.S $ddddg ptq H.uq]QP\RDE @@PR H S $ddddg ptq Huq]QP\RDEPR H&xVS $ddddg ptq H&Uuq]QP\RDEPR HP:cS $ddddg ptq HP:`auq]QP\RDEPRLQlS $ddddg ptqLQ0iuq]QP\RDEPR HHeuS $ddddg ptq HHequq]QP\RDEPR8Dus܁S $ddddg ptq8Dxs|uq]QG SymboldKd L$ch*f n4&0&/MGXI: created by Micrografx, Inc. Mgxgre v1.00   &0C (`tt`" " "" F" %P$P$  ??   @0@ @     ???]  A A @A A @A   @  @?!@ <x > ?? @?  |?pp@|  `x8?<  # ?? ? G 0?  @?D @~$ xx` Ā "8@  8`0@ 8p?   G@~8`  < @p ?0 00!@"! B`pB  ``@ B    8@  @@0?@@@@! 8@0B     08?@@ D< `@@`2 xa p! l b ?H@ G`O    @ # @?N @x@ (  H  0  ?B / @ !` B " @!@,D @pO x &    0  ( 8 8< r  `*     @@ (  8  P 8` P A@0 A  !8<     @ @D    H @X H0 $  8`< ` 0 0 `@ @ @ ( @ @ `@ @@ 0?" &@ @D @L            8<    @@@ @@@    `@ P@ P@ 0@ 8 <            8"< " " " @B DD       ? `             8@< @  @               8< @ @@ @@ @ @ @ @ @ @ @   8 <     D       ? `           8<               8<       @ @ @@ @@ @@ @@8@@<                 @ @ @( A  A A @A A @A ( A  A A @A A @A ?@0 ~`@?  " @`    @ @"  D   @ @"  @D   @ @" ?<@D |? @!@"  B@@  @ @"@1#0C 00 @!@ A )b(8801 | p&'<&,@ @  @  `@`0   --  "System- j .1 !L& N& & MathType`Symbol- 2 F%g 2 %g 2 %g 2 %g 2 U %g 2 %g 2 %g 2 %g 2 a%gSymbol- 2 Fk= 2 k= 2 k= 2 k= 2 U k= 2 k= 2 k= 2 k= 2 ak=Times- 2 F0 2 F9 2 0 2 8 2 0 2 7 2 0 2 6 2 U 0 2 U 5 2 0 2 0 2 3 2 0 2 2 2 a0 2 a1 2 FR.` 2 R.` 2 R.` 2 R.` 2 U R.` 2 R.4` 2 R.` 2 R.` 2 aR.` &  "System-'  --%?-- $--%3-- $38--%-- $}}--%-- $oo--%-- $--%7-- $7<--%O-- $4OY4--%Cf-- $OfuO--%z~-- $tt-- j (k)=-d 0 {}=-Ee(k)/d j (k)=d 0 {}Symbol- FMicrosoft Formel-Editor 2.0 DS Equation Equation.2NI  .1  &w & MathTypexSymbol-2 7{xSymbol-2  }xSymbol-2  {xSymbol-2  }Times- 2 FE 2 e 2 k 2 k 2 /E 2 e 2 k 2 kTimes- 2 ( 2 ) 2 /k 2 >( 2 ) 2 s( 2 ) 2 /k 2 '( 2 v)Symbol- 2 pd 2 ^ d 2 Yd 2 `d Symbol- 2 nj 2 WjSymbol- 2 ^ = 2  - 2  = 2 >- 2 G= Times- 2 j 0p 2 l0p &  "System-L ))))~EgCgC Ee(k)/d j (k)=d 0 {}@-4pd 0 2d m j mm=-N2N2  simd 0 2()2d m =4p"F e j (k+1) (m) * "F e j (k) (m)"F e j (k+1) (N-m) * "F e j (k) (N-m){  for  0m<N2-N2m-1"F e j (k+1) (m)=1Nf e j (k+1) (n)exp-j2pnmN[] n=0N-1  and:"F e j (k) (m)=1Nf e j (k) (n)exp-j2pnmN[] n=0N-1  0d 0 p FMicrosoft Formel-Editor 2.0 DS Equation Equation.2L5'x))))5'u M  .1   $0&0# & MathTypexSymbol-2 7{xSymbol-2  }-TTK,K-LWL +.w/ c) *7c77U7Y!Times- 2 FE 2 e 2 k 2 k 2 d 2 Jjk 2 sik 2 sm 2 d 2 R 9m 2 R ~ m 2 N 2 9m 2 N 2 #m 2 & |+m 2 & )N 2 V,m 2 - m 2 N 2 fk 2 n 2 jk 2 mnm 2 N 2  and Times - 2 m 2 m 2 >m 2 3N 2 FN 2 m 2 gec 2 kc 2 ec 2 kc 2 3ec 2 31kc 2 3ec 2 3kc 2 [ec 2 kc 2  ec 2 tkc 2 np 2 N 2 Iec 2  kcTimes- 2 ( 2 ) 2 /k 2 ( 2 m) 2 ~ 2 * ~ 2 cE~ 2 R ( 2 R X) 2 cx~ 2 R ( 2 R !) 2 ~ 2 !( 2 X) 2 x~ 2 ( 2 '$) 2 9~ 2  ( 2 L ) 2 d( 2 ) 2 exp 2 :k 2 v'~ 2 e ( Times - 2 \(J 2 *p 2 (J 2 m)J 2 3(J 2 & *p 2 3(J 2 3m)J 2 P(J 2 (J 2 > (J 2  )JSymbol- 2 Hd 2 g d 2 pd 2 0d 2 p 2 p Symbol- 2 Jj`Symbol- 2 j` 2 j` 2 k;j` 2 kj` 2 3j` 2 3~j` 2 j`Symbol- 2 F = 2 % @ 2 _- 2 ` 2 ` 2 ` 2 ` 2  2  2  2   2   2   2 % = 2 D` 2 R ` 2 - 2 ` 2 !- 2   2  2  2  2 I  2 & G* 2 B(- 2 !+ 2 - 2 .- 2 % = 2 ` 2 \` 2 - 2 $ 2 $ 2 R$ 2 p! 2 p! 2 Rp! Symbol- 2 ={ 2 -{ 2 <+{ 2 3+{ 2 0 +{ 2 +{ 2 ={ 2 H-{Symbol- 2  2  Times- 2 w 0p 2 0p 2 BE2p 2 9X2p 2 @0p 2 1)pJ 2 3.1)pJ 2  1)pJ 2 q1)pJ 2 0p 2 1pTimes - 2 24 2 2 2 _4 2 H)2 2 /1 2 1 2 U2Symbol- 2 R F% 2 R MF% 2 F% 2 MF% 2 F% 2 eF%Times -2 $ for ```` 2 & ,)0 2 & ,< 2 /G.N 2 Q q.2cTimes- 2 e- m 2 N 2 e(fk 2 en 2 eljk 2 nnm 2 pN Times - 2 ec 2 kc 2 np 2 NTimes- 2 eL ) 2 e( 2 eM) 2 eexp Times - 2 (J 2 y)JSymbol- 2 e% = 2 e` 2 eX- 2  2  2  2   2   2   2 !!$  2 !!W Symbol- 2 ={ 2 -{Symbol- 2  Times - 2 n1 2 e2 2 !! 0 Times- 2 *0p 2 I1p 2 !P0p`Symbol- 2 j`Symbol- 2 ep 2 !!@d 2 !!xp &  "System-2 m 2 xm  Π_T&SL(S F e j (k+1) (m)=f e j (k+1) (n)exp-j2pnmN[] n=0N-1  and:F e j (k) (m)=f e j (k) (n)exp-j2pnmN[] n=0N-1 z FMicrosoft Formel-Editor 2.0 DS Equation Equation.2   .1   @& & MathType-AEmqSymbol - 2 z7F% 2 E 7F% Times- 2 ec 2 kc 2  ec 2 F kc 2 tnp 2 tN 2 ec 2 kc 2 ec 2 \ kc 2 np 2 NTimes۠- 2 zFm 2 z| fk 2 zn 2 zjk 2 Ynm 2 N 2 ;and 2 E \m 2 E fk 2 E n 2 E jk 2 Onm 2 o N`Symbol- 2 j` 2 R j` 2 j` 2 h j`Symbol- 2 zp 2 E /p Times۠- 2 q(J 2  (J 2 q(J 2 O)J 2 (J 2 )JTimes- 2 z( 2 ze) 2 z.( 2 z) 2 zexp 2 :k 2 E ( 2 E {) 2 E Z ( 2 E ) 2 E exp Symbol- 2 O+{ 2  +{ 2 ={ 2 ( -{ 2 ={ 2 >-{Symbol- 2 z6= 2 z:` 2 z- 2  2  2 3 2 \ 2 \ 2 3\ 2 E L= 2 E f` 2 E - 2 B 2 | B 2 B 2  2 |  2 Symbol- 2 f 2 | Times- 2 1)pJ 2 ? 1)pJ 2  0p 2 1p 2 0p 2 1pTimes- 2 zG2 2 E s2 &  "System-- 2 `7F%L h _T&Sx(S F  "de(k)  (m)=4pNF e j (k+1) (m) * F e j (k) (m)(J 2  FMicrosoft Formel-Editor 2.0 DS Equation Equation.2"   .1  &@{ & MathType-bI Symbol2- 2 @7F% 2 @6 F% 2 @yF%`Times- 2 ~V 2 x(5 2 ')5Times- 2 @( 2 @[) 2 @( 2 @d) 2 @( 2 @) Timesǟ- 2 p (J 2 *p 2 (J 2 )J`Symbol- 2 dO 2  j` 2 j`Symbol- 2 J9p`Timesß- 2 9eG 2 kGTimes- 2 @<m 2 jN 2 @Em 2 @m Timesß- 2  ec 2  kc 2 ec 2 kcSymbol- 2 @,= 2 @ ` 2 @` Symbol- 2 N +{Times- 2 J~4 Timesß- 2  1)pJ &  "System-"System-L"PX _Tt&S(S f "d e (k) (n)=4pNf e j (k+1) (l) l=0N-1  f e j (k) (l+n)Times- FMicrosoft Formel-Editor 2.0 DS Equation Equation.2%V ? .1  @&` & MathType-Times- 2 xfk 2 n 2 N 2  fk 2 lk 2 fk 2 &lk 2 "n`Times- 2 g~eG Times- 2 /Lkc 2  ec 2 Y kc 2  l> 2 N 2 Uec 2 kc Times- 2 ~x 2 /(J 2 /)J 2  (J 2 F(J 2 $)JTimes- 2 J( 2 ) 2 A( 2 @) 2 ( 2 ) Symbol- 2 /dnSymbol- 2 yp`Symbol- 2 e j` 2 j`Symbol- 2 l= 2 ` 2 ` 2 + Symbol- 2  +{ 2  ={ 2 ; -{Symbol- 2 y Times- 2 4 Times- 2 R1)pJ 2 } 0p 2 1p &  "System-2 R1)pJ 2 } 0L%d _TX&S'S f "d e (k) (n)=4pNf e j (k+1) (-n)*f e j (k) (n) FMicrosoft Formel-Editor 2.0 DS Equation Equation.2d".]  .1   &z & MathType-Times- 2 @xfk 2 @n 2 j:N 2 @ fk 2 @n 2 @Nfk 2 @n`TimesA- 2 '{eG Times- 2 Hkc 2 K ec 2  kc 2 ec 2 kc TimesA- 2 U~x 2 (J 2 )J 2 : (J 2 (J 2 )JTimes- 2 @D( 2 @) 2 @} ( 2 @) 2 @( 2 @a) Symbol- 2 dnSymbol- 2 Kp`Symbol- 2  j` 2 !j`Symbol- 2 @{= 2 @` 2 @ - 2 @(* Symbol- 2  +{Times- 2 K4 TimesA- 2  1)pJ &  "System-Ld" X _T&S'S f "d e (k) (x)=f e j (k+1) (-x)*f e j (k) (x) FMicrosoft Formel-Editor 2.0 DS Equation Equation.28   .1  @&f & MathTypeTimesL- 2 xfk 2 fk 2 wfk`      !"#$%'+,-./13456789:;<=>?@ABCDEFGHIKOQSTUVWXYZ[\]_cefghijklmnoprvwxyz{|}Times- 2 g{eG TimesL- 2 /Hkc 2 uec 2 kc 2 ec 2 =kc Times- 2 ~x 2 /(J 2 /)J 2 d(J 2 (J 2 )JTimesL- 2 D( 2 ) 2  ( 2  ) 2 9( 2 ) Symbol- 2 /dnSymbol- 2 x 2  x 2 x`Symbol- 2 j` 2 Jj`Symbol- 2 z= 2 5 - 2 Q * Symbol- 2 A +{ TimesL- 2  1)pJ &  "System-o |,o ,o +o L8  @_TSdS f "d e (k) (x)fk FMicrosoft Formel-Editor 2.0 DS Equation Equation.2[ f .1  `& f & MathTypeTimesOP- 2 xfk`Times'6- 2 g~eG TimesOP- 2 /Lkc Times'6- 2 ~x 2 /(J 2 /)JTimesOP- 2 J( 2 ) Symbol- 2 /dnSymbol- 2 x &  "System- L[ @_TSS f "d e (k) (n)=f "d e (k) (x n =nj 0 )where:j 0 =4pN  is the sampling interval 2 g~e FMicrosoft Formel-Editor 2.0 DS Equation Equation.2R ~   .1   & & MathType`-Times'- 2 ixfk 2 in 2 ifk 2 iG n2 @where 2 N`Times- 2 P~eG 2 PeG Times'- 2 Lkc 2 kc 2  np Times- 2 ~~x 2 (J 2 )J 2 ~W~x 2 O(J 2 - )JTimes'- 2 iJ( 2 i) 2 i ( 2 i) 2 @<:k Symbol- 2 dn 2 JdnSymbol- 2 i& x 2 ij 2 vCj 2 ypSymbol- 2 il= 2 i = 2 iI` 2 vl= Times'- 2 0p 2 j0pTimes- 2 42 v is the s``k`k`2 v ampling in+kk`k2 v7tervalkk &  "System-LR  @_TSdS f "d e (k) (x) FMicrosoft Formel-Editor 2.0 DS Equation Equation.2  f .1  `& f & MathTypeTimes- 2 xfk`Times'- 2 g~eG Times- 2 /Lkc Times'- 2 ~x 2 /(J 2 /)JTimes- 2 J( 2 ) Symbol- 2 /dnSymbol- 2 x &  "System-L[ @E_T  f "d e (k) (n)&0 y+ .1  `&  & MathTypeTimes- 2 `xfk 2 `n`Times- 2 G~eG Times- 2 Lkc Times- 2 u~x 2 (J 2 )JTimes- 2 `J( 2 `) Symbol- 2 dn+&LMathTypeUU@ f "d e (k) (n) &  "System-  L[~ EgCgC d m @1NF  "de(k)  (m)F  "de(k)  (N-m){  for  0m<N2-N2m-1where:F  "de(k)  (m)=f "d e (k) (n)exp-j2pnmN[] n=0N-1 mp2[] FMicrosoft Formel-Editor 2.0 DS Equation Equation.2} y .1  &( & MathType-[}[5b Wb [Times - 2 ;d 2 N 2 ~ m 2  N 2 ; m 2 Sm 2 RN 2 Hmm2 ` ;where 2 <m 2 r fk 2 n 2 jk 2 onm 2 N Times- 2 m 2 tF kc 2 :jnp 2 3 jN`Times - 2 w eG 2 Oj kG 2 SeG 2  kG 2 eG 2 kG 2 x eGSymbol- 2 G@ 2 $` 2  - 2  2  2 s 2  2  2 u 2  2 Si 2 Ho- 2 H@ 2 H 2 H- 2 ,= 2 P` 2 - 2 , 2 , 2 ~, 2 r 2 r 2 ~r Symbol- 2 :={ 2 3 -{Symbol- 2 \Times - 2 1 2 r2 2 H1 2 ]2 Times- 2 :x0p 2 3 1pSymbol- 2 ~F% 2 F% 2 7F%`Times- 2 E ~V 2 O (5 2 O )5 2 a~V 2 (5 2 t )5 2 (~V 2 x(5 2 ')5Times - 2 ~T ( 2 ~ ) 2  ( 2 Z) 2 ` :k 2 ( 2 [) 2 D ( 2 ) 2 exp Times- 2 ~x 2 t (J 2 t )J`Symbol- 2 O; dO 2 dO 2 dO Symbol- 2 t dnSymbol- 2 pTimes-2  for ```` 2 SV0 2 S< 2 ]QN 2 }{2 &  "System-` 2  %` 2 C # L} EGS'Ul'U d m =14pf "d e (k) (v) -2p2p  exp-jmv2[]dv( FMicrosoft Formel-Editor 2.0 DS Equation Equation.2  .1  `& Y & MathType-=}=d==Times- 2 ;d 2 fk 2 M v 2 Ijk 2 !mv 2 dv Timeso- 2 m 2 O kc`Times- 2 eGSymbol- 2 G= 2 ` 2  ` 2 5- 2  2  2 Y 2  2  2 Y Symbol- 2 CX-{Symbol- 2 &Times- 2 1 2 4 2 2 Timeso- 2 C2p 2 2pSymbol- 2 Tp Symbol- 2 Obdn 2 CMp{ 2 p{ Times- 2 o~x 2 Og (J 2 OE )JTimeso- 2  ( 2  ) 2 ] exp &  "System-CRR2 ;dL ~EgClgC Ee(k)/d j (k)=d 0 {}=-2pd m expjmx2[] m=-  p-d 0 d 0 +p  dx=-2pd m  p-d 0 d 0 +p  expjmx2[] m=-  dx=-4pd m 1mexpjmp2[] m=-  2sinmd 0 2()=-4pd 0 d m j mn=-  simd 0 2()=-4pd 0 d 2m -1() mn=-  simd 0 () FMicrosoft Formel-Editor 2.0 DS Equation Equation.2L3-3-  .1   )&( & MathTypexSymbol-2 7{xSymbol-2 ; }-rro   }  G$ (__D >Symbol-2 X(>Symbol-2 X)Symbol-2 l(Symbol-2 l)")Times- 2 FE 2 e 2 k 2 k 2 d 2 jk 2 m 2  d 2 d 2 jk 2 $m 2  d 2 d 2 Jm 2 jk 2 + m 2  c$m 2 d 2 djk 2 sik 2 m 2 Nd 2 Nsik 2 Nm Times - 2 5m 2 Jam 2 {ym 2 m 2 xm 2 m 2 "m 2 m 2 7np 2 m 2 Mm 2 npTimes- 2 ( 2 ) 2 /k 2 2( 2 ) 2 exp 2 exp 2 exp 2 !sinkSymbol- 2 hd 2  d 2 *p 2 x 2  x 2 *p 2 $x 2  x 2 )p 2 + {p 2  @&d 2 )pd 2 ~d 2 N)pd 2 Np d Symbol- 2 5dj 2 xp{ 2 xdn 2 dn 2 p{ 2 Wp{ 2 ydn 2 ?Tdn 2 ?p{Symbol- 2 g = 2 T = 2 - 2 d` 2 ` 2 ! 2 .! 2 l! 2  2 . 2 l 2 T = 2 - 2 d` 2 $` 2 A  2 t  2   2 A 2 t  2  2 T = 2 - 2 c` 2 #` 2 + ` 2 i  2 V 2  2 i  2 V 2  2 v ` 2  %` 2 C # 2 *# 2 O # 2 C ( 2 *( 2 O ( 2 T = 2 - 2 ` 2 ` 2 ` 2 ` 2  2  2  2 N  2 N  2 N  2 NT = 2 N- 2 N` 2 N` 2 N6- 2 Ne` 2 N` Symbol- 2 J={ 2 J-{ 2 C> 2 x-{ 2 +{ 2 -{ 2 ?g+{ 2 ={ 2 H-{ 2  2 ={ 2 G-{ 2  2 7 ={ 2 7-{ 2 0P 2  ={ 2 -{ 2 PSymbol- 2 , 2 [ 2  2 rf 2 w e 2  2  Times- 2 5 0p 2 | J'0p 2 "0p 2 0p 2 0p 2 {2p 2 z!0pTimes - 2 n2 2 2 2 n2 2 E 2 2 n4 2 + 1 2 J2 2 2 2 J%2 2 n4 2 2 2 Nn4 2 N 1`Times- 2 0P 2 10P 2 0P 2 w0P &  "System- 2 N 2 e(fk ΠWTSlS f "d e (k) (x)=d m expjmx2[] m=-  FMicrosoft Formel-Editor 2.0 DS Equation Equation.2q  .     !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSUYZ[\]^_`abcdefghijlnrsuwxyz{|}~1  @& & MathType-Times۟- 2 xfk 2 A d 2 4jk 2  m`Times- 2 g~eG Times۟- 2 /Lkc 2  m 2 m Times- 2 ~x 2 /(J 2 /)JTimes۟- 2 J( 2 ) 2 / exp Symbol- 2 /dnSymbol- 2 x 2 xSymbol- 2 k= 2 f` 2 k  2 k  2 k  2  2  2  Symbol- 2 S={ 2 -{ 2 }Symbol- 2 Times۟- 2 2 &  "System-L WTSx S Ee(k)/d j (k)=d 0 {}=-2pF 0 (d 0 -p)=-2pF "d e (k) (d 0 +p)+F "d e (k) (d 0 -p)+F "d e (k) (d 0 -3p) =0 B-1[]=-2pF "d e (k) (d 0 +p)-F "d e (k) (p-d 0 )[]=-2pf "d e (k) (x) p-d 0 d 0 +p  dxF%  FMicrosoft Formel-Editor 2.0 DS Equation Equation.27Bq  .1  2&1z & MathTypexSymbol-2 7{xSymbol-2 ! }Symbol-2  [Symbol-2  N$]Times۟- 2 FE 2 e 2 k 2 k 2 F 2 rF 2 BF 2 &F 2 AF 2 F 2 +fk 2 $0d`Times- 2 eG 2 eG 2 a'eG 2 t eG 2 t `eG 2 t ,eG Times۟- 2 Okc 2 Oakc 2 O/(kc 2 < `kc 2 < .kc 2 < n-kcTimes- 2 ( 2 ) 2 /k 2 >( 2 ) 2 C( 2 ) 2 ( 2 4) 2 _( 2 $) 2 -)( 2 .) 2 ^( 2 ) 2 ,( 2 #) 2 l.( 2 /) Times۟- 2 :~x 2 O2(J 2 O)J 2  ~x 2 O(J 2 O)J 2 &~x 2 O'(J 2 O()J 2 ~x 2 < (J 2 < )J 2 ~x 2 < (J 2 < )J 2 ,~x 2 < -(J 2 < -)JSymbol- 2 pd 2 w d 2 p 2 d 2 p 2 p 2 d 2 1p 2 d 2 "p 2 )d 2 -p 2 p 2 d 2 p 2 p 2 "d 2 'p 2 .x 2 0x Symbol- 2 nj 2 O-dn 2 Odn 2 O&dn 2 < dn 2 < dn 2 < ,dn 2 0)p{ 2 0*dn 2 )dn 2 @+p{Symbol- 2 ^ = 2 % = 2 W- 2 $` 2 - 2 % = 2 W- 2 $` 2 + 2 + 2 !- 2 $+ 2 +- 2 T/- 2  2   2 + 2  2  1 2  1 2 + 1 2  1 2 % = 2 W- 2 $` 2 + 2 - 2 - 2 %= 2 B&- 2 )` Symbol- 2 *={ 2 0M*-{ 2 *+{Symbol- 2  m* Times۟- 2  0p 2 0p 2 0p 2 0p 2  0p 2 *0p 2 *0p 2 0p 2 #0pTimes- 2 *2 2 *2 2 ,3 2 _01 2 *2 2 '2`Times۟- 2 he+0P 2 >*0PMT Extra- 2  &1 2 )2 2 z'4 2 (4 2 r-3~ 2 0+4 2 ,4 &  "System- L7B@8 Р7VoU oU F 0 (x)=f 0 (u)du= -px  f "d e (k) (u+2p)du+ -px  f "d e (k) (u)du+ -px  f "d e (k) (u-2p)du -px  =f "d e (k) (u)du+ +px+2p  f "d e (k) (u)du+ -px  f "d e (k) (u)du -3px-2p  =F "d e (k) (x+2p)-F "d e (k) (p)+F "d e (k) (x)-F "d e (k) (-p)+F "d e (k) (x-2p)-F "d e (k) (-3p)=F "d e (k) (x+2p)+F "d e (k) (x)+F "d e (k) (x-2p)-1  FMicrosoft Formel-Editor 2.0 DS Equation Equation.2L0 0n    .1  ,&+d & MathTypeTimes- 2 FF 2 ofk 2 u 2 e du 2  fk 2 u 2 idu 2 fk 2 u 2 2du 2 Y"fk 2 %u 2 5*du 2 fk 2 B u 2  du 2 fk 2  u 2 Ydu 2 fk 2 Ru 2 du 2 F 2 F 2 qF 2 F 2 0F`Times- 2 eG 2 eG 2 _#eG 2 eG 2 eG 2 eG 2 eeG 2 eG 2 eG 2 ;eG 2 ~eG Times- 2 _kc 2 dkc 2 -$kc 2 kc 2 kc 2 kc 2 O 3kc 2 O kc 2 O kc 2 O kc 2 FOkc Times- 2 < 0p 2 <0p 2 2p 2 3p 2 &2pTimes- 2 $2 2 '2 2 2 2 ( 2 ) 2 ( 2 ) 2 ]( 2 ) 2 b( 2 ) 2 +%( 2 )) 2  ( 2  ) 2 ( 2 ) 2 ( 2 ) 2 1( 2 ) 2 ( 2 c) 2 ( 2 ) 2 ( 2 k!) Times- 2 ~x 2 (J 2 )J 2  ~x 2 (J 2 )J 2 "~x 2 #(J 2 $)J 2 k~x 2 c(J 2 A )J 2 4~x 2 ,(J 2  )J 2 {~x 2 s(J 2 Q)J 2 ~x 2 O (J 2 O )J 2 ~x 2 O (J 2 O g)J 2 9~x 2 O 1(J 2 O )J 2 ~x 2 O (J 2 O )J 2 ~x 2 F(J 2 F)JSymbol- 2 1x 2 p 2 (p 2 x 2 p 2 `p 2 x 2 h p Symbol- 2 p{ 2 Vxn 2  dn 2  p{ 2 r xn 2 dn 2 p{ 2 wxn 2 "dn 2 }!p{ 2 @!xn 2 ^dn 2 p{ 2 xn 2 p{ 2 'dn 2 p{ 2 xn 2 ndn 2 Tp{ 2 xn 2 p{ 2 O dn 2 O dn 2 O ,dn 2 O dn 2 FdnSymbol- 2 = 2 > = 2 + 2 C+ 2  + 2 &- 2 = 2 j + 2 3+ 2 = 2 + 2 - 2 5+ 2 - 2 - 2 >+ Symbol- 2  -{ 2 ( -{ 2 --{ 2  -{ 2 +{ 2 +{ 2 T-{ 2 i-{ 2 -{Symbol- 2 bB 2 b^  2 bc 2 b,! 2 q 2 q 2 qTimes- 2 M( 2  ) 2 ( 2 ) 2 1( 2  ) 2 ( 2 8) 2 c( 2 ) Times- 2 ~x 2 F(J 2 F)J 2 ~x 2 =(J 2 =)J 2 ~x 2 =(J 2 =i)J 2 ~x 2 =(J 2 =)JSymbol- 2 x 2  p 2  p 2 x 2  p 2 bx 2 x 2 p Symbol- 2 Fdn 2 =dn 2 =dn 2 =dnSymbol- 2  - 2  - 2 - 2 = 2  + 2  + 2  + 2 - 2 -Times- 2  2 2 a3 2  2 2 .2 2 1Times- 2 F 2 F 2 F 2 FF`Times- 2 ~4eG 2 ueeG 2 ueG 2 ueG Times- 2 Fkc 2 =3kc 2 =kc 2 =ekc &  "System-рg\*f.f Ee(k)/d j (k)=d 0 {}=-d 0"System- FMicrosoft Formel-Editor 2.0 DS Equation Equation.2[  .1  &w & MathTypexSymbol-2 7{xSymbol-2 ! }Times- 2 FE 2 e 2 k 2 kTimes- 2 ( 2 ) 2 /k 2 >( 2 )Symbol- 2 pd 2 w d 2 %d Symbol- 2 njSymbol- 2 ^ = 2 % = 2 W- Times- 2  0p 2 10p &  "System-L      ! "/#$%&'()*+,-.0I123456789:;<F>?@ABCDEGHJKLMjOPQRSTUVWXYZ[\]^`abcdefghiklmnopqrs}uvwxyz{|~ g'f.f G4_0 FMicrosoft Formel-Editor 2.0 DS Equation Equation.2=4o P .1   & & MathTypeP & L=4@EGS'Ux 'U F 0 (x)=1x+p2p0{ if xpxab<px-p9&Ba(( FMicrosoft Formel-Editor 2.0 DS Equation Equation.2 ' .1  @&@ & MathType--     Times- 2 FF Timeso- 2 ` 0pTimes- 2 1 2 *2 2 r0 2 ( 2 )Symbol- 2 1x 2 x 2 p 2 *_p 2  x 2  p 2  x 2 p 2 @, x 2 @LpSymbol- 2 = 2 + 2   2 A  2   2   2   2   2  < 2 @L  2 @y -Timeso- 2 K  if `k` &  "System-L рg'f+f f 0 (x)=12prectx2p[]Sy FMicrosoft Formel-Editor 2.0 DS Equation Equation.2o  .1  @ &`  & MathType-  Times- 2 xfk 2 rectk Times- 2 0pTimes- 2 1 2  2 2  2 2 ( 2 )Symbol- 2 x 2 p 2 [ x 2  pSymbol- 2 = 2 $` 2 8  2 8  2 8  2   2   2   &  "System-Lрg`'f+f Ee(k)/d j (k)-p,p(){}=0 2 FE FMicrosoft Formel-Editor 2.0 DS Equation Equation.2I  .1  &w & MathType>Symbol-2 f (>Symbol-2 B)xSymbol-2 7{xSymbol-2 }Timesg- 2 FE 2 e 2 k 2 kTimes- 2 ( 2 ) 2 /k 2 >( 2 ) 2  ,`Symbol- 2 pd 2  p 2 ? p Symbol- 2 njSymbol- 2 Q  2  - 2 =Times- 2 0 &  "System-? p L,  _T0'S(S f 0 (x)=d(x)      F 0 (x)=10.50{ if x>0x=0x<0 FMicrosoft Formel-Editor 2.0 DS Equation Equation.2q  [ .1  `& y & MathTypeTimesk- 2 xfk 2  F 2 ifkk Times- 2 0p 2  0pTimesk- 2 1 2 \0 2 n5 2 0 2 ( 2 ) 2 ( 2 ) 2 M ( 2  ) 2 .`Symbol- 2 x 2 d 2 ;x 2  x 2 x 2 x 2 xSymbol- 2 = 2 n = 2  2  2  2 Y 2 wTimesk-2  `````` 2 4 ` 2 j ` 2 !> 2 M0 2  = 2 L0 2 !< 2 M0 &  "System-Lq 8     #$%&(*+,-./0123456789:;<=>?@ABCDEGKLNPQRSTUVWXYZ[\]_cdegijklmnopqrstuvwxyz{|}@@g  d j (k)h{B + .1  @& i  & MathTypepSymbol- 2 `#d Symbol- 2 !jTimes- 2 `( 2 `@)Times- 2 `}k+&LMathTypeUU@ d j (k) &  "System-Lh{h@@g  d j (k)h{B + .1  @& i  & MathTypepSymbol- 2 `#d Symbol- 2 !jTimes- 2 `( 2 `@)Times- 2 `}k+&LMathTypeUU@ d j (k) &  "System-Lh{hg(f)f Ee(k)/d j (k)=p{}=-2p1-F 0 (0)()=-pand:Ee(k)/d j (k)=-p{}=p) 2  FMicrosoft Formel-Editor 2.0 DS Equation Equation.2. m .1   & & MathTypexSymbol-2 7{xSymbol-2 l }Symbol-2 (Symbol-2 )xSymbol-2 N7{xSymbol-2 NS }Times- 2 FE 2 e 2 k 2 k 2 F 2 `;and 2 FE 2 e 2 k 2 kTimes_- 2 ( 2 ) 2 /k 2 >( 2 ) 2 ( 2 c) 2 `:k 2 ( 2 ) 2 /k 2 >( 2 )Symbol- 2 pd 2 | p 2 1p 2 p 2 pd 2 c p 2 up Symbol- 2 nj 2 rnjSymbol- 2 ^ = 2 p = 2  - 2 o` 2 e- 2 = 2 - 2 ^ = 2  - 2 W =Times_- 2 u2 2 j1 2 0 Times- 2 g0p &  "System-L*рTSlS F 0 (p)=1F 0 (0)=0.5F 0 (-p)=0pd FMicrosoft Formel-Editor 2.0 DS Equation Equation.2M6   .1   &X & MathTypeTimes- 2 MFF 2 FF 2 FF Times- 2  0p 2  0p 2 P 0pTimes- 2 M1 2 @0 2 0 2 5 2 0 2 M( 2 M4) 2 ( 2  ) 2 .` 2 ( 2 )Symbol- 2 M1p 2 pSymbol- 2 M= 2 = 2 E- 2 = &  "System-- 2 LMt$V'U(U Ee(k)/d j (k)=-d 0 {}=-Ee(k)/d j (k)=d 0 {}xx\\~kP  FMicrosoft Formel-Editor 2.0 DS Equation Equation.2   .1  &w & MathTypexSymbol-2 7{xSymbol-2  }xSymbol-2  {xSymbol-2  }Times- 2 FE 2 e 2 k 2 k 2 /E 2 e 2 k 2 kTimes- 2 ( 2 ) 2 /k 2 >( 2 ) 2 s( 2 ) 2 /k 2 '( 2 v)Symbol- 2 pd 2 ^ d 2 Yd 2 `d Symbol- 2 nj 2 WjSymbol- 2 ^ = 2  - 2  = 2 >- 2 G= Times- 2 j 0p 2 l0p &  "System-LMxx\\LL  W&'W4*'W .1 FMicrosoft Formel-Editor 2.0 DS Equation Equation.2=4  P .1   & & MathTypeP &  L=4@ WD''W|*'W d 0 FMicrosoft Formel-Editor 2.0 DS Equation Equation.2Wh  .1   & & MathType`SymbolW- 2 `#d TimesS- 2 /0p &  "System- ULW,T `W&'W*'W F 0 (-x)=1-F 0 (x)Symbol- FMicrosoft Formel-Editor 2.0 DS Equation Equation.2 {  Z .1  @ &@  & MathType`Times3- 2 FF 2 F Times- 2  0p 2 0pTimes3- 2 1 2 ( 2 ) 2 r ( 2  )Symbol- 2 E- 2 = 2 -Symbol- 2 x 2  x &  "System-L {0h @W&'W)'W F 0 (x)  FMicrosoft Formel-Editor 2.0 DS Equation Equation.2"{   .1  @& & MathType`Times- 2 FF Times- 2  0pTimes- 2 ( 2 )Symbol- 2 1x &  "System-L"{Xh `W&'W)'W f 0 (x)=f 0 (-x)- 2 x FMicrosoft Formel-Editor 2.0 DS Equation Equation.22 {  H .1  @@ &  & MathType`Times- 2 xfk 2 -fk Times- 2 0p 2 0pTimes- 2 ( 2 ) 2 S( 2 )Symbol- 2 x 2 xSymbol- 2 = 2 - &  "System-F L2 {h`93x 3 Ee(k)/d j (k)=d 0 {}=x -d 0 -p-p   f 0 (x+2p)dx+x -p-d 0 +p   f 0 (x)dx=u-2p() -d 0 +pp   f 0 (u)du+x -p-d 0 +p   f 0 (x)dx=x -pp   f 0 (x)dx-2pf 0 (u) -d 0 +pp   du=-2pF 0 (p)-F 0 (p-d 0 )()=-2pF 0 (d 0 -p) FMicrosoft Formel-Editor 2.0 DS Equation Equation.2p.e   .1   *&)< & MathTypepxSymbol-2 "7{xSymbol-2 "! }>Symbol-2 W<      !#')+,-./012489:;<=>?@ABCDEFHJNPRSTUVWXY[_`abdfghijklmnopqrstuvwxyz{|}~(>Symbol-2 W)Symbol-2 (Symbol-2 )Times- 2 FE 2 e 2 k 2 k 2 0fk 2 @d 2  fk 2 r$d 2 Mu 2 MWfk 2 Mu 2 M%du 2 Mu!fk 2 MI%d 2 ]fk 2 1d 2 fk 2 u 2 du 2 OF 2 IF 2 X#FTimes- 2 ( 2 ) 2 /k 2 >( 2 ) 2 V( 2 ) 2 !( 2 #) 2 M}( 2 M) 2 M"( 2 M#) 2 ( 2 ) 2 7( 2 ) 2 ( 2 =) 2 ( 2 ]) 2 $( 2 l))Symbol- 2 pd 2 w d 2 7x 2 x 2 p 2 x 2 px 2 >"x 2 2%x 2 Mp 2 MGx 2 M#x 2 M &x 2 lx 2 x 2 x 2 p 2 p 2 :p 2 4p 2 d 2 l!p 2 >%d 2 i(p Symbol- 2 Fnj 2 dn 2 vp{ 2 vp{ 2 p{ 2 dn 2 zp{ 2 dn 2 wp{ 2 P-p{ 2 p{ 2 Pdn 2 PQp{ 2 Jp{ 2 ap{ 2 Jfdn 2 Jp{ 2 p{Symbol- 2 ^ = 2 % = 2 P` 2 + 2 ` 2 (+ 2 ` 2 #` 2 M% = 2 M- 2 Mw` 2 M` 2 M+ 2 M ` 2 M$` 2 % = 2 }` 2 ` 2 - 2 <` 2 < ` 2 % = 2 W- 2 $` 2 - 2 - 2 = 2 - 2 "` 2 L'- Symbol- 2 Y-{ 2 -{ 2 -{ 2 ]-{ 2 \-{ 2 +{ 2 Y-{ 2 +{ 2 4-{ 2 P3-{ 2 P+{ 2 JY-{ 2 J-{ 2 J|+{Symbol- 2 lZ 2 l 2 [ 2 j 2 - 2 - Times- 2 F 0p 2 F0p 2 F!0p 2 0p 2 !0p 2 0p 2 0p 2 0p 2 0p 2 0p 2 $0p 2 J&0p`Times- 2 u0P 2 !x0P 2 ( u0P 2 O0P 2 0PTimes- 2 #2 2 M2 2 B2 2 *2 2  2 &  "System-Lp.eT @W)V*V 0<d 0 pp &  FMicrosoft Formel-Editor 2.0 DS Equation Equation.2qW   .1   & & MathType`Times- 2 `70 Times- 2 k0pSymbol- 2 `L< 2 `fSymbol- 2 `_d 2 `p &  "System-)V 4NR@QLqW8T W8&'W''W Ee(k)/d j (k)=d 0 {}=x -d 0 -pp   f 0 (x)dx+x p-d 0 +p   f 0 (x-2p)dx=x -d 0 -pp   f 0 (x)dx+u+2p() -p-d 0 -p   f 0 (u)du=x -pp   f 0 (x)dx+2pf 0 (u) -p-d 0 -p   du=2pF 0 (-d 0 -p)-F 0 (-p)()=2pF 0 (-d 0 -p)=2pF 0 (d 0 ab-p)) 2 (  FMicrosoft Formel-Editor 2.0 DS Equation Equation.2L/I3 @W<(V+V -p<d 0 0 2 `0  FMicrosoft Formel-Editor 2.0 /I&  9 .1    +&* & MathTypexSymbol-2 .7{xSymbol-2 .! }>Symbol-2 c7(>Symbol-2 c)Symbol-2 (Symbol-2 )-Times˦- 2 FE 2 e 2 k 2 k 2 0fk 2 d 2 bfk 2 p$d 2 Y0fk 2 Yd 2 Yu 2 YT!fk 2 Y"u 2 Y"%du 2 ]fk 2 1d 2 Gfk 2 u 2 !du 2 uF 2 |F 2 r#F 2 KFTimes- 2 ( 2 ) 2 /k 2 >( 2 ) 2 V( 2 ) 2 ( 2 #) 2 YV( 2 Y) 2 Yz"( 2 Y#) 2 ( 2 ) 2 m( 2 ) 2 ( 2 p) 2 ( 2 Q) 2 $( 2 m*) 2 Ki( 2 K|)Symbol- 2 pd 2 w d 2 7x 2 x 2 x 2 4x 2 x 2 "p 2 0%x 2 Y7x 2 Yx 2 Yx 2 Yp 2 lx 2 x 2 x 2 p 2  p 2 Bd 2 mp 2 Np 2 !p 2 ?&d 2 j)p 2 K p 2 Kd 2 Kyp Symbol- 2 Rnj 2 dn 2 vp{ 2 ,p{ 2 ^p{ 2 dn 2 >p{ 2 dn 2 vp{ 2 \,p{ 2 p{ 2 \dn 2 \=p{ 2 cp{ 2 ap{ 2 ckp{ 2 hdn 2 p{Symbol- 2 ^ = 2 % = 2 P` 2 a` 2 + 2 ` 2 ) - 2 #` 2 Y% = 2 YP` 2 Ya` 2 Y+ 2 Y+ 2 Yt ` 2 Y$` 2 % = 2 }` 2 ` 2 + 2 >` 2 r ` 2 % = 2 J` 2 t- 2 P- 2 A- 2 {- 2 = 2 "` 2 q%- 2 M(- 2 K% = 2 KJ` 2 K\- Symbol- 2 Y-{ 2 -{ 2 -{ 2 +{ 2 Y-{ 2 -{ 2 -{ 2 \ -{ 2 \-{ 2 cY-{ 2 c-{ 2 -{ 2 ~-{Symbol- 2 xZ 2 xW 2 Z 2 V 2 F 2 F Times- 2 R 0p 2 R0p 2 R0p 2 0p 2 !0p 2 0p 2 0p 2 <0p 2 N0p 2 C0p 2 9$0p 2 K'0p 2 0p 2 "0p`Times˦- 2 u0P 2 -<0P 2 4 u0P 2 <0P 2 0PTimes- 2 S!2 2 Y2 2 D2 2 P2 2  2 2 KP2 &  "System-DS Equation Equation.2gWVQ  .1   &` & MathType`SymbolW- 2 `<- 2 `^< 2 `xSymbol- 2 `p 2 `qd TimesW- 2 }0pTimesv- 2 `0 &  "System-  LgWT~EgCx gC Ee(k)/d j (k)=d 0 {}=x -d 0 -p-d 0 +p  f e(k)/d j (k) (x/d 0 )dx FMicrosoft Formel-Editor 2.0 DS Equation Equation.2! t .1  &@ & MathTypexSymbol-2 7{xSymbol-2 ! }Times- 2 FE 2 e 2 k 2 k 2 1fk 2 d Times - 2 Mec 2 Mzkc 2 M"kcTimes- 2 ( 2 ) 2 /k 2 >( 2 ) 2  ( 2 /k 2 D) Times - 2 M(J 2 M)J 2 MT/> 2 M(J 2 M)JSymbol- 2 pd 2 w d 2 8x 2 x 2 d 2 bx Symbol- 2 @nj 2 dn 2 wp{ 2 dn 2 wp{ 2 Mdn`Symbol- 2 .j`Symbol- 2 ^ = 2 % = 2 Q` 2 ` Symbol- 2 Z-{ 2 -{ 2 Y-{ 2 +{Symbol- 2 f[ Times- 2 @ 0p 2 @0p`Times - 2 v0P 2 ;u0P &  "System-L!  1FDesigner 3.1 Zeichnung Designer Designer !31Symbol Newman.D.D`3>TimesicatsatedSansic3 "Systemwf$?{!   H !SLL i2dd,dS < E 2 %%SxS| 2%%SxX8 2Sxx@X 2S  H( 2S@ @ @H( 2S  H( 2S  H( 2SX X XH( 2S  H( 2SJ*~^n#SimU" 2fS113gq1 0S@@\guU" /(SgyU" /)S[[n}U) 7xS\\FTfn#S*gU" /2S))nU) 7pS\\TMfn#S /nU) 7-SBBlgU" /2SkknU) 7pS \ \8 U) 7xS\\}U) 7pS`\`\?fn#S /nU) 7-S??nU) 7pShh< 2S@ 2S2Jf zn#SgU" / S-iU" 2fS1111Jg1 0SgU" /2SXXugU" /(SOOlgU" /)SssnU) 7xSznU) 7pSnU) 7-S@@ 2S   2S"Jfn#SiU" 2fS116g1 0SgU" /2SCC`gU" /(S;;XgU" /)S^^nU) 7xSfnU) 7pSnU) 7+S@@( 2SX@X@ 2<ST$T$ =#Sj,j,{; P[S,,{; P]S\Z^ S@@U" 2fS@@U" 2fSr@r@U" 2fSu@u@ U" 2fS @ @ U" 2rectSmm21 eSJmJm`1 kSmm1 kS5o5oF1 (Sfofow!1 )Szozo%1 /Soo)1 (Soo-1 )SCC71U" /(StCtC5U" //SCC9U" /)SCC=U" /(SCCAU" /)SCCEU" /(SCCIU" /)SCCMU" /(SCCQU" /)SjjU1  dSY# jS5;5;]U) 7xS;;aU) 7dS;;\eU) 7xS;;iU) 7pS;;/mU) 7xS;;2qU) 7xSv;v;uU) 7pSnyU) 7xS11n}U) 7dS"}"}U) 7pSll1 0Sll1 0Sll1 0Sll1 0Sm1m1g1 0SvCvCU" /2SLCLCvU" /2S"U" /2S+C+CZU" /=SCCU" / SCCU" / SCCLU" /+S]C]CsU" / S9;9;gU) 7-S(;(;VU) 7+S;;=U) 7+S;;U) 7S&nU) 7+StU) 7SU) 7SXXU) 7StU) 7SU) 7SXXU) 7SL|L| $2SRe#S\Z^ S@@U" 2rectSnU) 7xS''nU) 7dS}}{U) 7pSnU) 7+StU) 7S U) 7SXX U) 7St U) 7S U) 7SXX U) 7Sc1c1|g 1 0S U" /2S   2Sf.nn#S .n U) 7-S..n! U) 7dSi1i1g% 1 0Sh 2PSDDLL\ 2SLF&)  2 dd,d,m  &  I &Creator: Mgxwmf 2.0 05/27/92&$Authors: Bob Giese and Scott Vollmer&Compatible with PSCRIPT -m- $'cr'-%#C- $C%rr $\,r,\,%cr%rc-%%%__%KK%%TimesSystNm;Ti-2 f Timesm.mu -2 0Timesm.mu -2 .(Timesl.lu -2 Z)Symbol^mAb&H&H(- 2 >xTimes&Hv@.l8Ti- 2 /2Symbol^mAm E0F&H- 2 Hp2 K-Times.l8m E0F&H- 2 r2- 2 p2 x2 p2 -2 p-%+3-%crcTimes.l 8m E0F&H- 2 / Times&HN@Nm;Ti-2 <f Times&H @.l Ti-2 L0Times&H?.m8Ti-2 22 c(2 )Symbol^mAb&H&H(-2 sx2 p2 -%c_c%ccTimes&HZ@Nl;Ti-2 f-2 0-2 T22 (2 )Symbol^mAb&H&H(-2 x2 mp2 0+-%Kc%#Symbol^l_b&H&H(-2 m[2 ]-%Times&H?Nm;Ti-2 f2 f2 f2 9f 2 )rect Times&H?Nm "Ti-2 e2 ,k2 k Times&H?.m Ti-2 (2 <)2 H/2 y(2 )Times&Hr?.l8Ti-2 (2 /2 -)2 (2 F)2 (2 )2 `(2 )Symbol^m&b&H&H(-2 Sd Symbol^m b&H&H(-2 ej-2 x2 d2 x2 $p2 x2 px2 p2 rx2 rd2 p Times&H?.l Ti-2 02 02 02 J02 0Times&H?.l8Ti-2  22 22 22 H=2  2 X 2 n+2  Symbol^mAb&H&H(-2 -2  +2 +2 2 r+2 v{2 {2 {2 v2 2 - %-%6u6Times&Hr?NmM;Ti- 2 &rectSymbol^lAb&H&H(-!2 x2 8d2 J/p2 +2 2 S2 42 w2 Sw2 4w Times&H?.m Ti-"2 \0-2 O2-%77+Symbol^mAb&H&H(-#2 7-2 7/d Times&H@.m Ti-$2 TR0%v-% $-%%'cr'%      !"#$L,^@WVT$V f 0 (x)sS- 2 xfk FMicrosoft Formel-Editor 2.0 DS Equation Equation.2{   .1  @&` & MathType`Times- 2 xfk Times- 2 0pTimes- 2 ( 2 )Symbol- 2 x &  "System-p) + f 0L{Dh W'VH)V f e(k)/d j (k) (x/d 0 )= f 0 (x-2p) + f 0 (x)+f 0 (x+2p)[]rectx+d 0 2p[]J 2  FMicrosoft Formel-Editor 2.0 DS Equation Equation.2,c  .1  `(& (2 & MathTypeSymbol-2 i [Symbol-2 E]-=}#=p'Times- 2 xfk 2 Y fk 2 fk 2 4fk 2 h rectk Times- 2 ec 2 kc 2 ukc Times- 2 f(J 2 D)J 2 /> 2 (J 2 )JTimes- 2 w( 2  /k 2 { ) 2  ( 2 ) 2 ( 2 ) 2 b( 2 ) Symbol- 2 dn`Symbol- 2 Ej`Symbol- 2 x 2 d 2 x 2 p 2 Fx 2 x 2 p 2 #x 2 %d 2 W%p Times- 2 0p 2  0p 2 0p 2 0p 2 &0pTimes- 2 *2 2 2 2 $2 2 M = 2   ` 2 t ` 2 + 2 > `Symbol- 2 - 2 + 2 + 2 ` 2 $+ 2 " 2 " 2 "" 2 ' 2 ' 2 "' &  "System-  p L,<W (V)V f e(k)/d j (k) (x/d 0 )dx=Pw:e(k,w) x,x+dx(]/d j (k,w)=d 0 {}=Pw:D j (k,w)-d j (k,w) x,x+dx(]/d j (k,w)=d 0 {}=Pw:D j (k,w) x+d 0 ,x+d 0 +dx(]/d j (k,w)=d 0 {}=f D j (k)/d j (k) (x+d 0 /d 0 )dxU"d  FMicrosoft Formel-Editor 2.0 DS Equation Equation.2L3 tX3 nc  .1   /&.~ & MathTypeSymbol-2 (Symbol-2 {]xSymbol-2 {xSymbol-2 i&}Symbol-2 (Symbol-2 G$]xSymbol-2 {xSymbol-2 5.}Symbol-2  (Symbol-2 #]xSymbol-2 -{xSymbol-2 --}Times- 2 xfk 2  d 2 P 2 <e 2 rk 2 d 2 Q k 2 P 2 k 2 >k 2 "d 2 (k 2 P 2 k 2 U"d 2 'k 2 Lfk 2 {d Times- 2 &ec 2 &kc 2 &ckc 2  Qkc 2  kc Times- 2 &`(J 2 &>)J 2 &/> 2 &(J 2 &)J 2  (J 2  )J 2  */> 2  (J 2  t)JTimes- 2 _( 2 /k 2 w ) 2 :k 2 ( 2 (,` 2 ) 2 ,` 2 A/k 2 ( 2 !,` 2 ") 2 :k 2 ( 2 _,` 2 ) 2 ( 2 ,` 2 ) 2 ,` 2  %/k 2 '( 2 (,` 2 {*) 2 :k 2 ( 2 _,` 2 ) 2 h,` 2 $/k 2 4'( 2 v(,` 2 *) 2 ( 2 /k 2 %) Symbol- 2 &dn 2 j 2 @Oj 2 @j 2 @&j 2 QOj 2 Qf&j 2  udn`Symbol- 2 ^pj` 2 V ^j` 2 V j`Symbol- 2 x 2 d 2  x 2 w 2 w 2 Lx 2 x 2 x 2 d 2 !w 2 $d 2 w 2 w 2 d 2 mw 2 x 2 a x 2 r#x 2 %d 2 L)w 2 ,d 2 w 2 w 2 x 2 d 2 x 2 !d 2 #x 2 j%d 2 (w 2 2,d 2 kx 2 d 2 ld 2 ;x Times- 2 0p 2 %0p 2 @-0p 2 Q0p 2 Q+ 0p 2 Q<-0p 2  0p 2  v0pSymbol- 2 . ` 2  = 2  2 + 2 #= 2  = 2 - 2 u 2 !+ 2 a+= 2  = 2  2 + 2  + 2 (!+ 2 += 2 = 2 + 2 ` 2 GD 2 GD Symbol- 2  D &  "System-@W&V*V d j (k)=d 0 FMicrosoft Formel-Editor 2.0 DS Equation Equation.2{f < .1  @& & MathTypepSymbolW- 2 `#d 2 `*d Symbol- 2 !jTimesW- 2 `( 2 `@)Times- 2 `}kSymbol- 2 `= Times- 2 60p &  "System-L{Lh`Wp(V*V e(k)=D j (k)-d j (k) FMicrosoft Formel-Editor 2.0 DS Equation Equation.2x{h l .1  @ &  & MathTypepTimes- 2 `3e 2 `ik 2 `k 2 ` kTimes- 2 `( 2 `,) 2 ` ( 2 `Z) 2 ` ( 2 `` )Symbol- 2 `= 2 `+- 2 `1D Symbol- 2 ;j 2 A jSymbol- 2 `C d &  "System-Lx{4h      !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSUY[\]^_`abcdefgimoqrstuvwxz~@WVT$V f 0 (x) FMicrosoft Formel-Editor 2.0 DS Equation Equation.2{   .1  @&` & MathType`Times- 2 xfk Times- 2 0pTimes- 2 ( 2 )Symbol- 2 x &  "System-p) + f 0L{Dh 1FDesigner 3.1 Zeichnung Designer Designer+d % B% &Creator: Mgxwmf 2.0 05/27/92&$Authors: Bob Giese and Scott Vollmer&Compatible with PSCRIPT - :!31Symbol NewmanXX`3>TimesicatsatedSansic3 "SystemV#?{!   H S@ 2S < E 2 %%SxS| 2%%S 2,X, Sxx 2,X, SXXx 2,X, SxX8 2Sxx@X 2S  H( 2S@ @ @H( 2S  H( 2S  H( 2SX X XH( 2S  H( 2SC^&n#SiU" 2fS113g1 0S@@\gU" /(SgU" /)S[[nU) 7xSTTTD 2dSlTlT 0L 2S\\F}&n#S*gU" /2S))nU) 7pS\\Tn#S /nU) 7-SBBlgU" /2SkknU) 7pS \ \8 U) 7xS\\}U) 7pS`\`\n#S /nU) 7-S??nU) 7pS4h4hd 2S] F ~#Si}U" 2fS:T:TJ~y# eSg@g@{zu1 kS  5Vq1 ~SRBRBaxm1 (SBBxi1 )SgeU" /(SgaU" /)S==Q}]1  dSnYU) 7xSddnUU) 7pS)nQU) 7-S::dgMU" /2SZ.n ~#SiIU" 2fS:T:TJ~E# eSg@g@{zA1 kS  5V=1 ~SSBSBbx91 (SBBx51 )Sg1U" /(Sg-U" /)S==Q})1  dSn%U) 7xSXX$n?.n ~#Si!U" 2fS:T:TJ~# eSh@h@|z1 kS  5V1 ~SSBSBbx1 (SBBx 1 )Sg U" /(SgU" /)S==Q}1  dSnU) 7xSggnU) 7pS+nU) 7+S>>hgU" /2Shh< 27Q- $cOc% - $ %OO $9,Oe,9,%cO;%O;(%vcO%cO%Oc-%%%;;%((%vv%ccTimesSystN7Q;Ti-2 f Timesn.nu -2 0Timesn.nu -2 (Timesm.mu -2 )Symbol^7QAbZYZY(-2 x- %W%vKTimesZYd.n8Ti- 2  2Symbol^7QAn E0FZY- 2 $p2 '-Times.n8n E0FZY- 2 N2- 2 gp2 x2 p2 -2 p%33TimesZYjNn;Ti- 2 yf TimeswNwu -2 e TimesoNou -2 k Timesd.du -2 ~2 (2 )Times7.7u -2 {(Timesw.wu -2 {)Symbol^d&bZYZY(-2 dSymbol^nAbZYZY(-2 vx2 v`p2 v"-TimesZY.o8Ti-2 {G2TimesZYNw;Ti-2 f TimesmNmu -2 e Times7QN7Qu -2 k Timesn.nu -2 ~2 (2 )Timesn.nu -2 (2 )Symbol^7Q&bZYZY(-2 dSymbol^nAbZYZY(-2 xTimesZYdNm;Ti-2 vf TimesdNdu -2 e TimesnNnu -2 k Times7.7u - 2 ~ Timesn.nu -!2 (2 )-2 (Timesd.du -"2 l)Symbol^7&bZYZY(-#2 d-2 x2 Jp2 +TimesZY.n8Ti-$2 22-%3      !"#$L+ @gFW  f "d e (k) (x)Nh + .1  `&} & MathTypeTimes- 2 xfk`Times- 2 g{eG Times- 2 /Hkc Times- 2 ~x 2 /(J 2 /)JTimes- 2 D( 2 ) Symbol- 2 /dnSymbol- 2 x+&LMathTypeUU@ f "d e (k) (x) &  "System-L[@WVT$V f 0 (x) FMicrosoft Formel-Editor 2.0 DS Equation Equation.2{]  .1  @&` & MathType`Times- 2 xfk Times- 2 0pTimes- 2 ( 2 )Symbol- 2 x &  "System-FL{Dh g'f)f f 0 (x)=f "d e (k) (x-2p)+f "d e (k) (x)+f "d e (k) (x+2p)[]rectx2p[] 2 F"x FMicrosoft Formel-Editor 2.0 DS Equation Equation.2`(>R  .1  @$&`$ & MathTypeSymbol-2 [Symbol-2 k]-!#Times/k- 2 xfk 2 fk 2 Hfk 2 fk 2 rectk`Times- 2 geG 2 gKeG 2 geG Times/k- 2 /kc 2 /kc 2 /lkc Times- 2 0pTimes/k- 2 Q 2 2 -2 2 !2 2 ( 2 ) 2 ( 2  ) 2 ( 2 d) 2 h( 2 ) Times- 2 =~x 2 /3(J 2 /)J 2 ~x 2 /(J 2 /)J 2 ~x 2 / (J 2 /)JSymbol- 2 x 2  x 2  p 2 x 2 x 2 p 2 F"x 2 "p Symbol- 2 /0dn 2 /dn 2 / dnSymbol- 2 = 2 + - 2  + 2 2+ 2 + 2 ` 2 #! 2 #! 2 #! 2 # 2 # 2 # &  "System-#'FI 3L`(@gFW  f "d e (k) (x)Nh + .1  `&} & MathTypeTimes- 2 xfk`Times- 2 g{eG Times- 2 /Hkc Times- 2 ~x 2 /(J 2 /)JTimes- 2 D( 2 ) Symbol- 2 /dnSymbol- 2 x+&LMathTypeUU@ f "d e (k) (x) &  "System-L[@WVT$V f 0 (x)F FMicrosoft Formel-Editor 2.0 DS Equation Equation.2{   .1  @&` & MathType`Times'- 2 xfk Times+- 2 0pTimes'- 2 ( 2 )Symbol- 2 x &  "System-p) + f 0L{DhU?Tx ?T f D j (k)/d j (k) (x/d 0 )   =f "D j (k)/d j (k) (x-2p/d 0 )+f D j (k)/d j (k) (x/d 0 )+f D j (k)/d j (k) (x+2p/d 0 ) []rectx2p[] = f 0 (x-d 0 -2p) + f 0 (x-d 0 )+f 0 (x-d 0 +2p)[]rectx2p[]@ FMicrosoft Formel-Editor 2.0 DS Equation Equation.2      $&()*+,-./1689:;<>CDEFGHIJLPQRSTUWYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~> ^  .1   9&8v & MathType0-57Symbol-2 0 [Symbol-2 0 $] ) *Times/- 2 pxfk 2 `fk 2 `+fk 2 `>#fk 2 `2rectk 2  fk 2  fk 2  `fk 2  &rectk Times- 2 kc 2 )kc 2  kc 2  kc 2 4kc 2 kc 2 G%kc 2 'kc Symbol- 2 D 2 ^D 2 D 2 #D`Symbol- 2 j` 2 5j` 2 <j` 2 < j` 2 @j` 2 j` 2 S$j` 2 &j` Symbol- 2 dn 2  dn 2 Zdn 2 m&dnSymbol- 2 px 2 pd 2 `x 2 `p 2 `.d 2 `Tx 2 `Ad 2 `g)x 2 `v,p 2 `.d 2 f^6x 2 6p 2  q x 2  d 2  p 2  $x 2  cd 2  x 2  ?d 2  5#p 2  )x 2 , )p Times/- 2 "(J 2 )J 2 [/> 2 (J 2 )J 2 uf~x 2  (J 2 f )J 2  /> 2 0 (J 2  )J 2 (J 2 )J 2 /> 2 }(J 2 [)J 2 $(J 2 %)J 2 !&/> 2 '(J 2 n()JTimes- 2 pE( 2 p/k 2 pi ) 2 ` ( 2 `p/k 2 `) 2 `( 2 `/k 2 ` ) 2 `(( 2 `-/k 2 `M0) 2  ( 2  ) 2  ( 2  ) 2  ( 2  8$) Times/- 2  0p 2 :0p 2 M 0p 2 /0p 2 b M0p 2 b 0p 2 b 0p 2 b o0p 2 b 0p 2 b K 0pTimes- 2 `X2 2 `+2 2 52 2  2 2  y"2 2 , ,)2 2 `L ``` 2 `0 ` 2 `{8 ` 2  = 2  q ` 2  2 ` 2  + 2  $ `Symbol- 2 `= 2 `.- 2 `+ 2 `!+ 2 `*+ 2  2 a 2  2 O1 2 aO1 2 O1 2 `'2` 2 ;5 2 ;5 2 ;5 2 7 2 7 2 7 2  - 2  - 2  K- 2  + 2  '- 2  N!+ 2  c%` 2 % w( 2 ^ w( 2 w( 2 % + 2 ^ + 2 + &  "System-0 0 128 128 128 240 [L> #XP J'K.K d 0|w FMicrosoft Formel-Editor 2.0 DS Equation Equation.2W  .1   & & MathTypePPSymbolJ- 2 #d Times- 2 40p &  "System-4g4xJ wLW,TP@J'K(K D j (k)  FMicrosoft Formel-Editor 2.0 DS Equation Equation.2   .1  ` &  & MathTypepPSymbolJ- 2 >D PSymbol- 2 HjTimesJ- 2 ( 2 _)Times- 2 k &  "System-Formatvorlage: MathematikScL|P J'K.K d 0|w FMicrosoft Formel-Editor 2.0 DS Equation Equation.2W  .1   & & MathTypePPSymbolJ- 2 #d Times- 2 40p &  "System-4g4xJ wLW,TP O!J  "D(k).   .1  `@& & MathTypePTimes- 2 M~ 2 2( 2 z)PSymbol- 2 >DTimes- 2 k&,MathTypeUU  "D(k) &  "System-e 2 kL|Ѡg<"f#f f "D(k)/d j (k) (x/d 0 )=f "d e (k) (x-d 0 -2p)+f "d e (k) (x-d 0 )+f "d e (k) (x-d 0 +2p)[]rectx-d 0 2p[]j FMicrosoft Formel-Editor 2.0 DS Equation Equation.2 3   .1  @`.& . & MathTypeSymbol-2  [Symbol-2 -]-GXTimes- 2 "xfk 2 "k fk 2 "-fk 2 ""fk 2 _2 rectk TimesC- 2 kc 2 kc 2 ?kc 2 kc 2 $kc`Times- 2 q eG 2 3eG 2 #eG TimesC- 2 7~x 2 (J 2 r)J 2 /> 2 <(J 2 )J 2 7 ~x 2  (J 2 )J 2 7~x 2 (J 2 )J 2 72#~x 2 *$(J 2 %)JTimes- 2 "( 2 "B/k 2 " ) 2 "=( 2 ") 2 "( 2 "u ) 2 "%( 2 "9-) Symbol- 2 D Symbol- 2 dn 2  dn 2 dn 2 %#dn`Symbol- 2 j`Symbol- 2 "x 2 "d 2 "x 2 "d 2 "p 2 "yx 2 "d 2 "&x 2 "@(d 2 "6,p 2 XOx 2 Xd 2 0p TimesC- 2  0p 2 0p 2 0p 2 L)0p 2 0pTimes- 2 ".2 2 "z+2 2 t2Symbol- 2 " = 2 "- 2 "- 2 "+ 2 "- 2 "G!+ 2 "('- 2 "O*+ 2 _ ` 2 Xv- 2 u  2   2   2 uo 2 o 2 o &  "System-L 3 (*@TSx S d j (k)Times FMicrosoft Formel-Editor 2.0 DS Equation Equation.2h{K  .1  @& & MathTypepSymbolD- 2 `#d Symbol- 2 !jTimes- 2 `( 2 `@)TimesW- 2 `}k &  "System-Lh{hP@O!7K  d e (k)EW[ + .1   &_ & MathTypePPSymbole- 2 #d Times- 2 4edTimese- 2 ckTimes- 2 ( 2 ')+&LMathTypeUU@ d e (k)F  &  "System-LEWlTP J (K$/K f "D(k)/d j (k) (x/d 0 )=f d e (k)/d j (k) (x-d 0 /d 0 )=f d e (k) (x-d 0 )2  FMicrosoft Formel-Editor 2.0 DS Equation Equation.2&& I .1  "&@"j & MathTypeTimesϞ- 2 sfl 2  fl 2 fl Times- 2 kd 2 kd 2  kd 2 +kd 2 _kd`TimesϞ- 2  eG 2 eG Times- 2 v~w 2 (J 2 j)J 2 /? 2 '(J 2 )J 2 9 (J 2 )J 2 i/? 2 (J 2 )J 2 (J 2 )JTimesϞ- 2 |( 2  /l 2  ) 2 $( 2 /l 2 U) 2 X( 2 !) PSymbol- 2 D PSymbol- 2 dp 2 5 dp 2 dp 2 dp`PSymbol- 2 ?j_ 2 ;j_PSymbol- 2 x 2 d 2 x 2 d 2 d 2 x 2  d Times- 2 0p 2 0p 2 0p 2 !!0pPSymbol- 2 Y = 2 - 2 #= 2 - &  "System-L&P J'K.K d 0|w FMicrosoft Formel-Editor 2.0 DS Equation Equation.2W  .1   & & MathTypePPSymbolJ- 2 #d Times- 2 40p &  "System-4g4xJ wLW,TP O!J  "D(k).   .1  `@& & MathTypePTimes- 2 M~ 2 2( 2 z)PSymbol- 2 >DTimes- 2 k&,MathTypeUU  "D(k) &  "System-e 2 kL|PJKK "D(k)=d j (k)+e j (k+1)-e j (k)[] 2pa =d j (k)+d e (k) FMicrosoft Formel-Editor 2.0 DS Equation Equation.2#qvj  .1   & & MathTypeaPSymbol-2 n [aPSymbol-2     "#$%&'()*+,-./01359;<=>?@ABCEIJKLMNPRSTUVWXYZ[\]^_`abcdefghijklmnopqrstvz|}~] -((Timesg- 2 M~ 2 2( 2 z) 2 ^( 2 ) 2 \ ( 2 ) 2 ( 2 ) 2 ( 2 Y) 2 ( 2 )PSymbol- 2 >DTimesg- 2 k 2 k 2  e 2  k 2 me 2 \k 2 k 2 Bk Times- 2 *edPSymbol- 2 g= 2 g+ 2  + 2 P- 2 = 2 +PSymbol- 2 d 2 Rd 2 %d PSymbol- 2 j 2  j 2  j 2 p{ 2 IjTimes- 2  1 Timesg- 2 j2p &  "System-YiToVoWoYoYpYqL#q`WVlV f d e (k) (x)=d(x-(u+2p)) -2p-p  f "d e (k) (u)du+d(x-u) -pp  f "d e (k) (u)du+d(x-(u-2p)) p2p  f "d e (k) (u)du=f "d e (k) (x-2p)+f "d e (k) (x)+f "d e (k) (x+2p)      if  x -p,p(]0                                                     else{=f "d e (k) (x-2p)+f "d e (k) (x)+f "d e (k) (x+2p)  []rectx2p[]k 2 @? k FMicrosoft Formel-Editor 2.0 DS Equation Equation.2L.c.c    .1  )&) & MathType@>Symbol-2 %(*Symbol-2 )]Times- 2 xfk 2  u 2 fk 2 u 2 ldu 2 7u 2 fk 2 "u 2 F$du 2  u 2 fk 2 ku 2 du 2 fk 2 6fk 2 fk 2 Hfk 2 fk 2 Ffk`Times+- 2 ?~eG 2 eG 2  eG 2 ieG 2 ~eG 2 ~<eG 2 ~eG 2 NeG 2 eG 2 LeG Times- 2 Lkc 2 Ikc 2 Ix!kc 2 1kc 2 Fn kc 2 F kc 2 Flkc 2  kc 2 kc Symbol- 2 dn 2 =p{ 2 ]p{ 2 I:dn 2 =p{ 2 p{ 2 I dn 2 % p{ 2 1p{ 2 1dn 2 F dn 2 Fdn 2 Fdn 2 dn 2 Tdn 2 dnSymbol- 2 x 2 d 2 G x 2 p 2 d 2 x 2 d 2 $ x 2 p 2 x 2 p 2 x 2 x 2 p 2 l#x 2 &p 2 (p 2  x 2  p 2 0x Times+- 2 (J 2 )J 2 G~x 2 I?(J 2 I)J 2 ! ~x 2 I!(J 2 I!)J 2 ~x 2 1(J 2 1j)J 2 ~x 2 F (J 2 F )J 2 ~x 2 F(J 2 F)J 2 ~x 2 F (J 2 F)J 2 ~x 2 (J 2  )J 2 a~x 2 Y(J 2 7)J 2 ~x 2 (JTimes- 2 J( 2 ) 2 ( 2  ( 2 )) 2 ( 2 ) 2 q( 2 ) 2 v"( 2 #) 2 ( 2 n ( 2 )) 2 ( 2 3) 2 l ( 2 ) 2 ( 2 X) 2 j( 2 ) 2  ( 2 ) 2 ( 2 )Symbol- 2 k= 2 n - 2 - + 2 R` 2 F+ 2 - 2 + 2 K - 2  - 2 k= 2 - 2 + 2 *+ 2 + 2 $ 2  2  2  2 } 2  2 k= 2  - 2 v+ 2 + Symbol- 2 =-{ 2 -{ 2 =z-{Symbol- 2   2  2 Times- 2 X2 2 32 2 7 2 2 72 2  0 2  2 Times+- 2 =2p 2 2pTimes-2  if ``````k`` 2 &- 2 ',`2  ``````````2 u ``````````2 5 ``````````2  ``````````2  ``````````2 u  else```kSymbol-2 [Symbol-2  ]-a$B& Times+- 2 kcTimes- 2 L!rectk Times+- 2 )JTimes- 2 ( 2 )Symbol- 2 x 2 p 2 $x 2 )2%pSymbol- 2 + 2  ` 2 "# 2 [# 2 # 2 "Y& 2 [Y& 2 Y&Times+- 2 2 2 )v$2 2 / `` &  "System-   W'WX'W f d e (k) (x)=f d e (k),"d e (k) (x,u) -2p2p  du=f d e (k)/"d e (k) (x/u) -2p2p  f "d e (k) (u)du=f d e (k)/"d e (k) (x/u) -2p-p  f "d e (k) (u)du+f d e (k)/"d e (k) (x/u) -pp  f "d e (k) (u)du                                                 +f d e (k)/"d e (k) (x/u) p2p  f "d e (k) (u)du- 2  FMicrosoft Formel-Editor 2.0 DS Equation Equation.2L/wx /w  .1  `+& +k & MathTypeTimes- 2 xfk 2 9fk 2 <u 2 ldu 2 fk 2 u 2 fk 2 2"u 2 #du 2 9fk 2 u 2 fk 2 Xu 2 du 2 fk 2 "u 2 $fk 2 6(u 2 )du 2 (fk 2 "u 2 $fk`Times- 2 S~eG 2 ? eG 2  eG 2 eG 2 eG 2 eG 2 ? eG 2 eG 2 eG 2 eG 2 eG 2 %eG 2 .eG 2 eG 2 %eG Times- 2 Lkc 2 ] kc 2 ]p kc 2 ]kc 2 ]qkc 2 ] kc 2 kc 2 kc 2 kc 2 kc 2 ukc 2 &kc 2 mkc 2 mkc Symbol- 2 dn 2 ]dn 2 ] dn 2 Qp{ 2 Tp{ 2 ]dn 2 ] dn 2 Qnp{ 2 .p{ 2 ]Ndn 2 dn 2 3 dn 2 y p{ 2 ]p{ 2 tdn 2 dn 2 dn 2 y ;p{ 2 p{ 2 R%dn 2 mdn 2 m"dn 2 ap{ 2 Cp{ 2 mc%dnSymbol- 2 x 2  x 2 x 2 x 2 x 2  x Times- 2 (J 2 )J 2 ] (J 2 ] ),J8 2  ~x 2 ] (J 2 ] )J 2 ](J 2 ]f)J 2 ]/> 2 ~x 2 ](J 2 ])J 2 [~x 2 ]S (J 2 ]1!)J 2 (J 2 )J 2 /> 2 @ ~x 2 8 (J 2  )J 2 ~x 2 y(J 2 W)J 2 (J 2 j)J 2 /> 2 ~x 2 (J 2 )J 2 _%~x 2 W&(J 2 5')J 2 m(J 2 m{)J 2 m/> 2 /~x 2 m'(J 2 m )J 2 p%~x 2 mh&(JTimes- 2 J( 2 ) 2 n ( 2 ,` 2 ) 2 o( 2 /k 2 ) 2 !( 2 ") 2 ( 2 >/k 2 ) 2 ( 2 ) 2 s ( 2 "/k 2 #) 2 '( 2 () 2  ( 2 -"/k 2 #)Symbol- 2 k= 2 E= 2 f` 2 k= 2 ` 2 + 2 j$` 2 + 2 {$` Symbol- 2 Q-{ 2 Qy-{ 2 y -{ 2 -{ 2 y -{Symbol- 2 4  2 4 2 \  2 \  2 D Times- 2 Q2p 2 2p 2 Q2p 2 2p 2 y 2p 2 2pTimes-2  ``````````2  ``````````2 T ``````````2  ``````````2  ````````` Times- 2 m&kcTimes- 2 G(u 2 )du Times- 2 mF')JTimes- 2 '( 2 )) &  "System-P@O!7K  d e (k)EW[ + .1   &_ & MathTypePPSymbole- 2 #d Times- 2 4edTimese- 2 ckTimes- 2 ( 2 ')+&LMathTypeUU@ d e (k)F  &  "System-LEWlT ΀W'Wl'W d e (k)="d e (k)+2p"d e (k) -2p,-p(]"d e (k) if"d e (k) -p, p(]"d e (k)-2p "d e (k)  p, 2p(]{ (u) FMicrosoft Formel-Editor 2.0 DS Equation Equation.2 a .1  &n & MathType>Symbol-2  (*Symbol-2 Z]>Symbol-2 J(*Symbol-2 W]>Symbol-2 (*Symbol-2 N]SymbolX- 2 @#d 2 d 2 q p 2 ?d 2 (p 2 Xp 2 @#d 2 @d 2 @p 2 @p 2 d 2 p p 2 d 2 p 2 Lp Times- 2 -ec 2 `ec 2 `Hec 2 ,ec 2 ec 2 ec 2 ecTimes- 2 @Uk 2 k 2 pk 2 @T k 2 @ ifkk 2 @k 2 k 2 kTimes- 2 @( 2 @) 2 ~ 2 i( 2 ) 2 U~ 2 ( 2 3) 2 ,` 2 79~ 2 @( 2 @ ) 2 7~ 2 @a( 2 @) 2 @,` 2 w~ 2 j( 2 ) 2 w~ 2 m( 2 ) 2 ,`Symbol- 2 @= 2  + 2  2 - 2 - 2 @t 2 @- 2  - 2  2  2  2 ( 2 N 2 Times- 2  2 2 l2 2  2 2 2 2 @  ` 2 @z ` 2 g ` 2  ` 2  ` &  "System- 2 yj` 2 yL@@g  "d e (k) E~b %+ .1  &N & MathType`Times- 2 :~ 2 ( 2 )Symbol- 2 $d Times- 2 @-ecTimes- 2 Uk+&LMathTypeUU@ "d e (k)kument% q2;; &  "System-  LEl W)'W)'W F e j (k) (e)=121+ep+1pgabsine() 1-gab 2 cos 2 e()  arccos-gabcose()()[]   if eabp=1  if e>p =0 if e<-p 2 zu" FMicrosoft Formel-Editor 2.0 DS Equation Equation.2o2p N  s .1  ` -&- & MathTypeP-6 N  $$>Symbol-2 j(>Symbol-2 j)>Symbol-2 '.(>Symbol-2 'z)w --`U`U`O$>Y Y >v!v!>Symbol-2 n$(>Symbol-2 %)Symbol-2 (Symbol-2 7&)U))U**Times- 2 zFF Times- 2  ec 2 [kc`Symbol- 2 gj`Symbol- 2 ze 2 ] e 2 > p 2  p 2 ` g 2 `Ne 2 g 2 e 2 zn g 2 z$e 2 z*e 2 zw,p 2 C e 2 s p 2 e 2 " p Times- 2 (J 2 )JTimes- 2 zY( 2 z) 2 `sink 2 icos2 z arccos 2 zu"cosSymbol- 2 zw= 2 z` 2 z + 2 z + 2 `s` 2 D- 2 ` 2 zr` 2 zf- 2 z!` 2  2  2 / 2  2 & 2 & 2 /& 2 & 2 z^+ 2 w= 2 [ > 2 # < 2 O -Times- 2 1 2 2 2 z 1 2 1 2 I1 2 1 Times- 2 @2p 2 q2pTimes-2 z`' if ```k`2 8 if ``k` 2 F  ` 2 t=2 0 if `k` &  "System-Lo2p |@`p`p gab=g(k)ab(  FMicrosoft Formel-Editor 2.0 DS Equation Equation.2     # !"$>%&'()*+,-./012=456789:;<?@ABCDEF{fT . .1  @`&  & MathType`-DHHD__DDSymbol - 2 Wg 2 (gSymbol- 2 =TimesS- 2 ( 2 e)Times- 2 k &  "System- L{h`lpp f e j (k) (e)=12p1-gab 2 1-gab 2 cos 2 e()1+gabcose()p-arccosgabcose()()[] 1-gab 2 cos 2 e()  []=1-gab 2 2p1-gab 2 cos 2 e()()1+gabcose() 1-gab 2 cos 2 e()  p2+arctangabcose() 1-gab 2 cos 2 e()  ()[][]~x 2 O FMicrosoft Formel-Editor 2.0 DS Equation Equation.2L[<T8" gflf F "d e (k) (x)=f "d e (k) (z) -x  dz= -[Symbol-2 (>Symbol-2 J)_ JJii>Symbol-2 ;(>Symbol-2 )##$$>Symbol-2 '(>Symbol-2  ))Symbol-2 56#(Symbol-2 5))ZSymbol-2 Z[ZSymbol-2 Z*]MMll>Symbol-2 #(>Symbol-2 2%)k --VIVIV%"* l  .l .P  P  >Symbol-2 ;(>Symbol-2 );Symbol-2 + (;Symbol-2 )  Yl Yxl x>Symbol-2 J(>Symbol-2 ) 3  3 >Symbol-2 (>Symbol-2 ) m 5-u 5C |-C    U!  u! " #V,l V,u-l u->Symbol-2 G0(>Symbol-2 1) ,3 , .3 .>Symbol-2 |2(>Symbol-2 3) *m 2*-u 2*C y*-C * * * R4 ) r4Times- 2 kxfk Times- 2 ec 2 Jkc`Symbol- 2 Tj`Symbol- 2 ke 2 p 2 L g 2 | g 2 ~e 2 ag 2 e 2 p 2 #g 2 @(e 2 dg 2 f$e 2 & g 2 p 2 g 2 e 2 pg 2 e 2 g 2 e 2  "p 2 m,g 2 0e 2 ,g 2 2e Times- 2 (J 2 )JTimes- 2 kL( 2 k) 2 1cos 2 Bcos2 iarccos 2 %cos 2 !cos 2 ncos 2 Qcos 2 cos2  `%arctank 2 N.cos 2 /cosSymbol- 2 kr= 2 k` 2 L - 2 < - 2  ` 2 k` 2 k+ 2 ` 2 +` 2 V- 2 <%` 2 *` 2 $- 2  ` 2  2  2  2  2 + 2 + 2 + 2 + 2  r= 2  - 2 ` 2 y - 2 ` 2  ` 2  + 2 ` 2 - 2 '` 2  !` 2  L$+ 2 -` 2 +- 2 $/` 2 T) 2  ) 2  ) 2 ) 2 T|4 2  |4 2  |4 2 |4 2 4C" 2 u C" 2 C" 2  C" 2 4?5 2 u ?5 2 ?5 2  ?5 2 \ 2 \ 2 \ 2 \ 2 S \ 2 5 2 5 2 5 2 5 2 S 5Times- 2 t@1 2 2 2 L 1 2 U 1 2 k1 2 =1 2  1 2 2 2 1 2  1 2 1 2 @ #2 2 *1 Times- 2 on2p 2  2p 2 R2p 2 62p 2 g:#2p 2 t2p 2  2p 2 2p 2 K2p 2  2p 2 H.2p 2  12p &  "System- einem Stapel nach links aus Nichtx  f e j (k+1) (u)f e j (k) (z+u) -  dudz=f e j (k+1) (u)f e j (k) (z+u -x  ) -  dzdu=f e j (k+1) (u)F e j (k) (x+u) -  du=F e j (k) (u)f e j (k+1) (u-x) -  du2  FMicrosoft Formel-Editor 2.0 DS Equation Equation.2/B } .1  +&@+k & MathTypepTimes- 2 FF 2 fk 2  z 2  dz      !"#$%&'()*+,-./026789:<>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ\`acefghijklmnopqrtxz{|}~ 2 fk 2 u 2 :fk 2 z 2 u 2 dudz 2 fk 2 P u 2 fk 2 7z 2 Su 2 dzdu 2 Ifk 2 } u 2 "F 2 (u 2 )du 2 F 2  u 2 fk 2 u 2 du`TimesC- 2 eG 2 " eG Times- 2 ekc 2  kc 2 6ec 2 6mkc 2 6ec 2 6kc 2 Aec 2 A kc 2 AWec 2 Akc 2 Aec 2 Akc 2 As#ec 2 A$kc 2 % ec 2 % kc 2 % \ec 2 % kc TimesC- 2 ~x 2 (J 2 )J 2 ~x 2  (J 2 o )J 2 6(J 2 6(J 2 6)J 2 A (J 2 AH(J 2 A&)J 2 A(J 2 Ad$(J 2 AB%)J 2 % (J 2 % ~ )J 2 % M(JTimes- 2 c( 2 ) 2  ( 2 $ ) 2 U( 2 ) 2 ( 2 x) 2  ( 2  ) 2 ( 2 ) 2 ( 2 E!) 2 %( 2 Q)) 2 ( 2 G ) 2 ( 2 ") Symbol- 2 dn 2 dn 2 xn 2 xn 2 xnSymbol- 2 x 2 ;&x 2 Lx`Symbol- 2 nyj` 2 nj` 2 yj` 2 yj` 2 yj` 2 y#j` 2 ] j` 2 ] j`Symbol- 2 = 2 A= 2 Z` 2 + 2 = 2  ` 2 -+ 2 = 2 "` 2 c'+ 2 = 2  ` 2 /- Symbol- 2 x-{ 2 xu-{ 2 6+{ 2 x?-{ 2 { 2 Ae +{ 2 y-{ 2 -{ 2  2 A+{ 2 -{ 2 ! 2 % ++{ 2 g-{ 2  Symbol- 2 [ 2 [ 2 [v 2 f 2 f 2 f 2 J  TimesC- 2 6f1)pJ 2 A 1)pJ 2 A 1)pJ 2 % 1)pJ &  "System-L/B08  @_T,SS f "d e (k) (x)=f e j (k+1) (-x)*f e j (k) (x)=f e j (k+1) (u)f e j (k) (x+u) -  du F*F FMicrosoft Formel-Editor 2.0 DS Equation Equation.2(   .1  %&$Y & MathTypeTimes- 2 xfk 2 fk 2 ^fk 2 fk 2 u 2 vfk 2 "u 2 6#du`Times'- 2 ~eG Times- 2 OLkc 2 Sec 2 kc 2 ec 2 (kc 2 Yec 2 kc 2 ec 2 @kc Times'- 2 ~x 2 O(J 2 O)J 2 D(J 2 (J 2 )J 2 J(J 2 (J 2 )JTimes- 2 J( 2 ) 2  ( 2  ) 2 &( 2 v) 2 ( 2 ) 2 >( 2 ") Symbol- 2 OdnSymbol- 2 x 2  x 2 x 2 x`Symbol- 2 9j` 2 94j` 2 9j` 2 9Lj`Symbol- 2 k= 2  - 2 5 * 2 G= 2 ` 2  + Symbol- 2 " +{ 2 (+{ 2 C{-{ 2 Symbol- 2 & Times'- 2  1)pJ 2 1)pJ &  "System-*L( PҀ7K&oJoJ "d e (k)=e j (k+1)-e j (k)- 2 4 FMicrosoft Formel-Editor 2.0 DS Equation Equation.2  .1  `& n & MathTypepTimes- 2 :~ 2 ( 2 ) 2 ( 2  ) 2 T ( 2 )Symbol- 2 $d Symbol- 2 @j 2 @ j Times- 2 @-ecTimes- 2 Uk 2 e 2 k 2  e 2  kSymbol- 2 = 2 ,+ 2  -Times- 2 8 1 &  "System-L P@O!7K  d j (k)O[h[ + .1  `& j & MathTypepPSymbol- 2 #d PSymbol- 2 !jTimes۠- 2 ( 2 K)Timess- 2 k+&LMathTypeUU@ d  j (k)"D"D &  "System-" n Lh|рg (f)f D j (k)=j(k+1)-j(k)[] 2pa +e j (k+1)-e j (k)[] 2pa [] 2pa =d j (k)[] 2pa +e j (k+1)-e j (k)[] 2pa [] 2pa =d j (k)+e j (k+1)-e j (k)[] 2pa [] 2pa] FMicrosoft Formel-Editor 2.0 DS Equation Equation.26  .1  1&`1 & MathTypeSymbol-2 G[Symbol-2 G?] -aSymbol-2 m[aSymbol-2 mJ]=^Symbol-2 M[^Symbol-2 Z]aSymbol-2 [aSymbol-2  ]na Pa aSymbol-2 C [aSymbol-2 ]nlPl^Symbol-2 M[^Symbol-2 ]||aSymbol-2 "[aSymbol-2 g-]n-P-^Symbol-2 ][^Symbol-2 w/] 0 0Symbol0- 2 ">D Symbol- 2 Hj 2 }p{ 2 j 2 j 2 p{ 2 Q p{ 2 -j 2  p{ 2 oj 2 8j 2 p{ 2 $p{ 2 j 2 $j 2 *j 2 .p{ 2 0p{Symbol- 2 ",j 2 "m j 2 S/d 2 SdTimes/- 2 "( 2 "g) 2 "0( 2 " ) 2 "q ( 2 ") 2 "( 2 " ) 2 "|( 2 ") 2 S( 2 SL ) 2 S?( 2 S) 2 S( 2 SW) 2 S( 2 S ) 2 S$( 2 S'() 2 S+( 2 S,)Times- 2 "k 2 "k 2 " k 2 ":e 2 "?k 2 "e 2 "k 2 Sk 2 S e 2 Sk 2 Se 2 Sk 2 S* k 2 SW#e 2 S\%k 2 S *e 2 S%,kSymbol- 2 "8= 2 "+ 2 "X - 2 "+ 2 "T+ 2 "- 2 S8= 2 S. + 2 S+ 2 Sg- 2 SH= 2 S!+ 2 Sq&+ 2 S(-Times- 2 " 1 2 "`1 2 S1 2 S}'1 Times/- 2 2p 2 2p 2 Q 2p 2  2p 2 2p 2 2p 2 +.2p 2 ;02p &  "System-L6P`O!7K  e j (k+1)-e j (k)| ) h R .1  ` &  & MathTypepTimes- 2 @e 2 Nk 2 e 2 k PSymbol- 2 j 2 7jTimes- 2 ( 2 ),` 2  ( 2 a )PSymbol- 2 Y+Times- 2 i1 2 / ` &  "System-L) T|P7KoJoJ D j (k)=y j (k+1)-y j (k)[] 2pa =j(k+1)+2e j (k+1)[] 2pa -j(k)+2e j (k)[] 2pa [] 2pa =j(k+1)[] 2pa +2e j (k+1)[] 2pa [] 2pa -j(k)[] 2pa +2e j (k)[] 2pa [] 2pa [] 2pa =j(k+1)[] 2pa -j(k)[] 2pa +e j (k+1)-e j (k)[] 2pa FMicrosoft Formel-Editor 2.0 DS Equation Equation.2L4\ΠSSx S "j=j[] 2pa =jn2p -p,p4~Z   .1  ` 0&/ & MathType aPSymbol-2 Q[aPSymbol-2 ] -]iaPSymbol-2 [aPSymbol-2 / ]] i aPSymbol-2 #[aPSymbol-2 +]]Q,iQ,..PSymbol-2 N[PSymbol-2 N ]o o aPSymbol-2 kN[aPSymbol-2 k*]&2PSymbol-2 N{[PSymbol-2 N]WWaPSymbol-2 k6"[aPSymbol-2 k&]&&2&((&"+2 "+PSymbol-2 [PSymbol-2  ]  PSymbol-2 {[PSymbol-2 ] W WPSymbol-2 7 Q[PSymbol-2 7 t]  PSymbol - 2 W>D PSymbol- 2 Hj 2 j 2 \ j 2 9p{ 2 j 2 j!p{ 2 )j 2 ,p{ 2 4/p{ 2  p{ 2 j 2 ep{ 2 ep{ 2 p{ 2 n#j 2 K'p{ 2 e)p{ 2 +p{ 2 O C p{ 2 O p{ 2  j 2  j 2 p{PSymbol- 2 Wj 2 W$j 2 j 2 j 2 0j 2 jTimes- 2 W( 2 Wr) 2 W^( 2 W ) 2 W1( 2 W) 2 W( 2 W) 2 ~ 2 W`( 2 W) 2 W%( 2 Wp&) 2 k(~ 2 W)( 2 WC+) 2  ( 2 a ) 2 }~ 2 [( 2 ) 2 ( 2 I) 2 }"~ 2 C$( 2 %) 2 9( 2 ) 2 ( 2 I) 2 ( 2 +) 2 ( 2 )Times- 2 Wk 2 Wy 2 Wk 2 W y 2 Wk 2 Wk 2 We 2 Wk 2 W%k 2 Wq(e 2 W*k 2 k 2 e 2 k 2 k 2 "e 2 $k 2 k 2 k 2 We 2 ek 2 *e 2 8kPSymbol- 2 W<= 2 W+ 2 W - 2 W]= 2 W'+ 2 W+ 2 W+ 2 W"- 2 W;'+ 2  2 k 2 ) 2 - 2 k- 2 )- 2 <= 2  + 2 9+ 2 + 2  2 4 2  2 < 2 4< 2 < 2 - 2 !!+ 2  2 4 2  2 "( 2 4"( 2 "( 2 r 2 r 2 |r 2 `* 2 `* 2 |`* 2 <= 2 + 2 g - 2 !+ 2 p+ 2 -Times- 2 W 1 2 W71 2 W 1 2  1 2 1 2 1 2 1 Times- 2 2p 2  2p 2 ,2p 2 .2p 2  2p 2 2p 2 e22p 2 2p 2 &2p 2 e)2p 2 V+2p 2 O 2p 2 O 2p 2 > 2p &  "System-(]  and:  "jabp. FMicrosoft Formel-Editor 2.0 DS Equation Equation.2tV  .1  &`Q & MathTypeSymbol-2 [Symbol-2  ] -~`>                     # $ & ( ) * + , - . / 0 1 2 3 4 5 6 7 9 Symbol-2 (*Symbol-2 ]-O+O+Times- 2 E~ 2 ,` 2 #:k 2 ~Symbol- 2  j 2 j 2 j 2 i p 2 %p 2 p 2 j 2 Xp Symbol- 2 (Ip{Symbol- 2 |= 2 m= 2  2  ` 2   2 R- 2 ? Times- 2 (2pTimes- 2  2Times- 2  n 2 andTimes- 2  `` 2  `` &  "System-5a eKuD>5a vKJpJ uD?5eK uD?5vKuD@5a eKuD@5a vK uDa a uD uDA5eK;     # !"$>%&'()*+,-./012=456789:;<?@ABCDEFIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmopqrstuwxyz{R**`+a+b+u+v+w+x+z+{+++++++++++,,,,,@,A,T,U,V,W,,,,,,,,,,,,,,,,,,,&-'-(-G-H-[-\- uD45vKJpJuD55a eKuD55a vK uD65eK uD65vK uD75eK uD75vKuDuD95a eKuD95a vKuD:5a e8KuD:5a vK uDa a 8\-]-^-------------------*.+.>.?.@.A................//////////ĽuD+5avK uD-5eK uD-5vK uD.5eK uD.5vKJpJ uD/5eK uD/5vK uD05eK uD05vKuD15a eKuD15a vK uDa a  uD35eKuD35avKuD uD45eK3////0 0 00.0/0B0C0D0E0000000000001 1 1 1 11111E1F1Y1Z1[1\1v1w11111111111111ú uD$5K uD$5vK uD%5eK uD%5vK uD'5eK uD'5vKuD(5a eKuD(5a vK uD)5eK uD)5vKJp uD*5eK uD*5vKuDa  uDa uD+5a eK61111d2e2f2y2z2{2|2~22222222222333333333333546474J4K4L4M4O4P4\4]4_4d4e4x4y4z4{44444뾵뭤 uD5vK uD5eK uD5vKuD5a eKuD5a vKuD5a eKuD5a vK uD 5eK uD 5vKuD"5a eKuD"5a vK uDa a uD uD#5eK uD#5vK74445555555555555555555H6I6J6]6^6_6`6b6c6o6p6r6u666666666666666677 777#7$7Ļ uD5vKuD5a erKuD5a vK uD5eK uD5vKuD5a eKuD5a vK uD5eK uD5vKUuD5a eKuD5avK uDa a uD uD5K8$7%7&7z7{77777777777777777~8888888888888888888888::: ; ; ; ;;;;;;ž㶭㥜㔋uD 5a eKuD 5avKuD5a eKuD5a vKuD5a eKuD5a vK uD5eK uD5vKuD5a eKuD5a vK uDa a  uD5eK uD5vKuD uD5eK6;I;J;K;^;_;`;a;c;d;p;q;s;;;;;;;;;;<<<<<<<< <[<\<]<p<q<r<s<u<v<<<<<<<<<<<<<<<<==>=Q=R=ƿ uD5vKuD5a eKuD5a vKuD5a eKuD5avKJ uD 5eK uD 5vKJpJ uD 5eK uD 5vKuDuD 5a eKuD 5a vK uDa a :R=S=T=n=o=p=======================> > >>>>>>H?I?J?r?s?t??????????@@@@@uD5avKuD5a eKuD5a vKJpJJxhuD5a eKuD5a vKuD5a eKuD5a vKuD5a eKuD5avK uDa a uD uD5eK7@@@@@@@@@@&A'A:A;AE?EKELENE~EEEEEEEEEEEEEEEEEEEEEFF FVFWFjFkFlFmFoFpF|F}FzH{HHHHHHHHHHH J JJ uD5KuD5avK uD5KuD5avK uD5eKuD5avKuDuD5a e>KuD5avKuD5a eKuD5avKa  uDa uD5a eKuD5avK9JJ J!J KK]LLzN{NNNQQQQTTUTXX/Y0Y7\8\D\W\X\\\m__``!`"`d`e`````````` a aLaMaNaPaQaRaTaUaaaaaaaaaa!b"b:b;bzb{b|b~bbbbbbbb uDܠa uDa^bC uDCcPcUVJpuD uD5KuD5avKObbbbbbbTcUcmcncccccccccccddd"du uDaC uDCuD^b R `  < = M  6 7 B y * + !H!5!N,IIIIIIN,IIIIIIN,IIIIII d  , d XQ dQ<<$+ b c d e n !/9  4OyO y !{#%-%&&U''(?(()))p**N,II!!L!!!!!!H!!!k!!!!!Z!B!!g!!!!!!!: !!p!V!|!!!R!v <<< d  ,&*`+++,,,--/001111222d223354_4444&5'5(555H6r66 777888!!!!!!!!)!;! !!!p# p# p#hp#h!!!o!C!W!!!hp#p#p#p#!!!!@!4 !R! !h!h!@!!H F#)8 959g99N:^:u::;I;s;"<E<[<<<<<n===>o>>r??@@%AOAAAABpBBBCCCD=D!!!+!+!!!!!"!!!h!!!8!!!!!!!!!H!! !!: !!!!!!!h! !!!!!!Y<<*=DDD$ENE~EEE FUFFzHHHHH/I0I J"J~KKKKK]L!!!!Y!!!!!!! # # ##!]     !!< ] F  F  ]LL{NNNQQTUTXXM[\[7\8\D\\>]]a^^m_`!d"d!!!!!!!!' !!!!!^!!!!!!!!!!<<<!  4.!  4.K@Normala0`0 Heading 1,H1 < Ua c0k$,`, Heading 2,H2 <Ua c$,`, Heading 3,H3 <Ua c,`, Heading 4,H4 <Ua c(`( Heading 5,H5<Ua "A@"Default Paragraph Font o z-HTML Tag ]^bc$o$Definition Term,DT]c4o4Definition List,DL0<<a c8o"8 Gleichunghp#ach0`2hList Bullet,ULF  4ha c"a"d            OU$p'*0H3N7o;>ACJY"aV*****) * *  |R=7 + e n !/9  4OyOy{ "-"##U$$%?%%&&&p''`((()))**,--...//d//0051_111'2(222H3r33 444555 656g66N7^7u778I8s8"9E9[99999n:::;o;;r<<==%>O>>>>?p???@@@A=AAA$BNB~BBB CUCCzEE0F GHHH]II{KKNNNQUQUUMX\X7Y8YDYY>ZZa[[m\]!a"a"a$*************************$44D**D**D**$***************C %Z(*\-/14$7;R=@B9EJbVWXYZ[\]^_`abcdefgh+ *8=D]L"dijklmnw>?YZg$dfgiKP`MB w + ` H\^ Pdfiv02PdfivG[]""""*"#####V$j$l$o$|$%*%,%/%<%E%Y%[%%%%& &:&N&P&&&&&&<'P'R'q'''''a(u(w(z((((())@)T)V)))))))))G*[*]**********+>+@+++++++,,,,,,, -.-B-D-----. . ..E.Y.[.v.........e/y/{/~/////0000061J1L1O1\1d1x1z111122222222I3]3_3b3o3333333344#4%4z44444444~555555555557 8 888J8^8`8c8p8888999\9p9r9u9999999=:Q:S:o:::::::::::: ;s<<<<<=====&>:><>?>L>>>>>>>>>>?q????????????@ @ @@4@H@J@@@@@@A(A*A-A:AAAAAA%B9B;B>BKBBBBBBBBBBCVCjClCoC|CzEEEEEE GG G]]] ^L^N^O^Q^^^:_z_|_}____m`````a"a333C333C3333C@3C33C@333333999:5::5:5::5::5:5:5:5::5::5::5:5:5:::5:::5:::::5:::5:::::5::5:5:::5::5::5:::5:::5:5:5:::5:5::5::5:5:5:5:5:5:5:::5::5:5:5:5:5:5:5:::33C33C33C3R`<=M6 7 B y * /899 H` ioOy4OyG_< = "-"]"^"_"b"""##U$$$$$$%?%E%]%%&:&Q&&&&&&&&&''<'S'p''`((()@)X)))))&*(*G*_******+B+++++++,-.-F------.E.].v....//d////0051_1d1|111&2(2222222H3r3333 44'4z4444~5 656u778I8s8889E9[9999<:U:n:::;o;;;;H<L<r<<====%>O>>>>?Y?Z?p?????@4@L@@@A=AAA$BNB~BBB CUCCzEE/F0F G"G HHZDZJZQZvZZZZZZZZZZZZ[[a[h[n[t[[[[[[[m\t\z\\\\\\\\$a;Thomas Werschlein&\\Jazz\root\home\luzi\tmp\PRIVATE .doc@Compaq PageMarq 15 PostScript\\Excalibur\pagemarq-psPSCRIPTCompaq PageMarq 15 PostScriptCompaq PageMarq 15 PostScriptg 3d 0 ̡*1 dCustom page 1BBCustom page 2BBCustom page 3BBCompaq PageMarq 15 PostScriptg 3d 0 ̡*1 dCustom page 1BBCustom page 2BBCustom page 3BB 1Times New Roman Symbol &Arial"qh  P)!< MACROBUTTON HtmlDirect   PRIVATE MACROBUTTON HtmlDirect *  PRIVATE src="http://www.geo.unizh.ch/rsl/fringe96/icons/esa-grey.gif" height=28 width=79 alt="esa" align=middle MACROBUTTON HtmlResImg INCLUDEPICTURE "C:/WIN95/TEMP/wia8b0/esa-grey.bmp" \* MERGEFORMAT    PRIVATE
MACROBUTTON HtmlDirect *FRINGE 96   PRIVATE MACROBUTTON HtmlDirect * PRIVATE src="http://www.geo.unizh.ch/rsl/fringe96/icons/rsl-grey.gif" height=28 width=47 alt="RSL" MACROBUTTON HtmlResImg INCLUDEPICTURE "C:/WIN95/TEMP/wia8b0/rsl-grey.bmp" \* MERGEFORMAT    PRIVATE MACROBUTTON HtmlDirect *  PRIVATE MACROBUTTON HtmlDirect    PRIVATE href="http://www.geo.unizh.ch/rsl/fringe96"MACROBUTTON HtmlResAnchor  PRIVATE src="http://www.geo.unizh.ch/rsl/fringe96/icons/home.gif" height=28 width=74 alt="home" border=0 MACROBUTTON HtmlResImg INCLUDEPICTURE "C:/WIN95/TEMP/wia8b0/home.bmp" \* MERGEFORMAT   PRIVATE src="http://www.geo.unizh.ch/rsl/fringe96/icons/space.gif" width=15 height=1 alt="" MACROBUTTON HtmlResImg INCLUDEPICTURE "C:/WIN95/TEMP/wia8b0/space.bmp" \* MERGEFORMAT    PRIVATE href="http://www.geo.unizh.ch/fringe-bin/search"MACROBUTTON HtmlResAnchor  PRIVATE src="http://www.geo.unizh.ch/rsl/fringe96/icons/search.gif" height=28 width=78 alt="search" border=0 MACROBUTTON HtmlResImg INCLUDEPICTURE "C:/WIN95/TEMP/wia8b0/search.bmp" \* MERGEFORMAT    PRIVATE MACROBUTTON HtmlDirect  Estimating the derivative of modulo-mapped phases   PRIVATE MACROBUTTON HtmlDirect * Otmar Loffeld  PRIVATE
MACROBUTTON HtmlDirect *Zentrum fr Sensorsysteme (ZESS) Universitt-GH-Siegen, Paul-Bonatz-Str. 9-11 57068 Siegen, Germany loffeld@wiener.zess.uni-siegen.de http://www.zess.uni-siegen.deChristoph Arndt  PRIVATE MACROBUTTON HtmlDirect *Zentrum fr Sensorsysteme (ZESS) Universitt-GH-Siegen, Paul-Bonatz-Str. 9-11 57068 Siegen, Germany arndt@nv.et-inf.uni-siegen.de http://www.nv.et-inf.uni-siegen.de/pb2/www_pb2Achim Hein  PRIVATE MACROBUTTON HtmlDirect *Zentrum fr Sensorsysteme (ZESS) Universitt-GH-Siegen, Paul-Bonatz-Str. 9-11 57068 Siegen, Germany hein@nv.et-inf.uni-siegen.de http://www.nv.et-inf.uni-siegen.de/pb2/www_pb2  PRIVATE MACROBUTTON HtmlDirect *Abstract Nearly all known phase unwrapping techniques try to unwrap the mapped phases by a sequence of differentiating, taking the principal value of the discrete derivative and integrating again. This procedure, conceptually appealing as it may appear, however, yields strongly biased phase derivatives and thus strongly biased phase estimates as the paper will show. This bias is a serious drawback of any differentiation of modulo mapped noise contaminated functions. The problem lies in the methodology rather than in the wrap around effect, as no sequence of linear and nonlinear operations may be altered without seriously affecting the result. It can be shown Ja:11  H$   $C (210543876;:9>=<A@?DCBGFEJIHMLKPONSRQVUTYXW\[Z_^]ba`edchgfkjinmlqpotsrwvuzyx}|{~UUVVVWWWXXXYYYZZZ[[[\\\]]]^^^___```aaabbbcccdddeeefffggghhhiiijjjkkklllmmmnnnooopppqqqrrrssstttuuuvvvwwwxxxyyyzzz{{{|||}}}~~~ J:11  H$   $C (210543876;:9>=<A@?DCBGFEJIHMLKPONSRQVUTYXW\[Z_^]ba`edchgfkjinmlqpotsrwvuzyx}|{~UUVVVWWWXXXYYYZZZ[[[\\\]]]^^^___```aaabbbcccdddeeefffggghhhiiijjjkkklllmmmnnnooopppqqqrrrssstttuuuvvvwwwxxxyyyzzz{{{|||}}}~~~ J loffeld-et-al/loffeld-et-al.html100644 23423 23424 101440 6220531634 15513 0ustar luziguest Estimation the derivative of modulo-mapped phases
esa FRINGE 96 RSL
home search

Estimating the derivative of modulo-mapped phases

Otmar LoffeldZentrum für Sensorsysteme (ZESS), Universität-GH-Siegen, Paul-Bonatz-Str. 9-11
57068 Siegen, Germany
loffeld@wiener.zess.uni-siegen.de
http://www.zess.uni-siegen.de
Christoph ArndtZentrum für Sensorsysteme (ZESS), Universität-GH-Siegen, Paul-Bonatz-Str. 9-11
57068 Siegen, Germany
arndt@nv.et-inf.uni-siegen.de
http://www.nv.et-inf.uni-siegen.de/pb2/www_pb2
Achim HeinZentrum für Sensorsysteme (ZESS), Universität-GH-Siegen, Paul-Bonatz-Str. 9-11
57068 Siegen, Germany
hein@nv.et-inf.uni-siegen.de
http://www.nv.et-inf.uni-siegen.de/pb2/www_pb2

Abstract

Nearly all known phase unwrapping techniques try to unwrap the mapped phases by a sequence of differentiating, taking the principal value of the discrete derivative and integrating again. This procedure, conceptually appealing as it may appear, however, yields strongly biased phase derivatives and thus strongly biased phase estimates as the paper will show. This bias is a serious drawback of any differentiation of modulo mapped noise contaminated functions. The problem lies in the methodology rather than in the wrap around effect, as no sequence of linear and nonlinear operations may be altered without seriously affecting the result. It can be shown mathematically, that computing the discrete derivative of noisy modulo-2 mapped phase yields estimates of the unambiguous discrete derivative, which are always biased towards lower absolute values. Thus phase slopes are always underestimated. It can be shown further that the bias clearly depends on the phase slope itself as well as on the coherence. The paper will present the theoretical analysis, and give some hints of how to circumvent the problem.
Keywords: Phase unwrapping techniques, phase slope estimation, biased phase estimates

Introduction

The determination of the unambiguous phase from noisy observations of complex angularly modulated signals is an unsolved problem in general, especially if phase and amplitude are mutually uncorrelated or even independent. This is clearly the case for a complex SAR interferogram. In terms of signal theory a SAR interferogram can be considered as a complex, simultaneously amplitude and phase modulated 2D signal with non Gaussian error statistics. Usually the wanted interferometric phase is obtained by a simple tan-1 operation delivering phase values within the principal interval (e.g. -, , depending on the kind of inverse function). These phases contain all the information needed for the generation of digital terrain elevation maps of observed areas but they do not contain that information in an unambiguous way as any absolute phase offset (an integer multiple of 2) is lost. Furthermore they are subject to phase noise coming from the superimposed amplitude noise in real and imaginary part of the InSAR image. The process of resolving these phase ambiguities is usually called phase unwrapping which in terms of signal theory is simply a two dimensional phase demodulation problem. Nearly all classical approaches to phase unwrapping, known from optical interferometry apply a sequence of differentiating, taking the principal value of the discrete derivative and integrating again along specified paths. The paper will show that such a sequence of operations always yields more or less badly biased estimates of the wanted derivative of the unambiguous phase. This especially applies to a combination with Linear Least Squares techniques which are commonly used to reduce stochastic phase errors. The paper will present the analysis, the numerical, and the analytical evaluation of the phase bias depending on the phase slope and on the signal to noise ratio. Finally some hints to circumvent the problem will be given.

1. Theory and Concepts

1.1 1D Stochastic Analysis

Let the observed phase obtained from a tan-1-operation form real and imaginary part of the interferogram, be related to the unambiguous phase by the following mapping:

(1)

where (k) is the true unambiguous phase at time or point k, is the phase error and the bracket indicates the operation of taking the principal value of the argument phase term in a way that:

(2)

As a result of equations 1 and 2 the observed phase always lies within the base interval (-,. This effect, arising anytime when the interferometric phase is computed by a tan-1() operation from quadrature and inphase component of an interferogram, is well known as the wrap around effect.

Nearly all known phase unwrapping techniques with the exception of /LoKr1, LoKr2/ try to unwrap the mapped phases by a sequence of differentiating, taking the principal value of the discrete derivative and integrating again.

The operation of forming the discrete derivative yields:

(3)

In the sequence of operations in equation 3 we have only used the fact that adding any integer multiple of 2 does not change the result of a modulo-2 operation. are the phase errors at points k and k+1, respectively, mapped into the base interval (-,. The stochastic properties of these errors, namely distribution density and second order moments are known. Now again using the same identities in equation 3 as before we can further write:

(4)

where the last equality has exploited that along normal integration paths the true phase variation's modulus is always smaller than . Normal refers to those integration paths which do not cross a cutline. is the true discrete phase derivative.

Equation 4 clearly expresses the error which we commit when forming the discrete derivative from modulo-2 mapped noisy data. If there were no phase error present the result would be totally correct, but since phase errors always occur in normal interferograms, we commit a systematic error when 'differentiating' modulo-2 mapped data. The phase derivative will be - depending on the noise - more or less badly biased and so will be the unwrapped phase if the biased derivative is integrated again.

The further investigation of the bias error will be organized as follows. In a first step we will investigate the stochastic features of the phase error difference in the second term of equation 4. Then we will evaluate, how a modulo operation, which occurs twice, changes the known distribution density of random variable. Let us introduce the non-mapped phase error difference variable by:

(5)

Obviously the numerical values can vary between 2, as any of the terms in the difference can vary between . Assuming that the phase errors of two subsequent phase samples are independent random variables with identical and symmetric distributions the resulting distribution density of the phase difference is the correlation product of the individual phase distribution densities. Thus we have:

(6)

Later on we will need the distribution function rather than the density. This function is simply obtained by:

(7)

Conceptually equation 7 can be solved since the individual terms are known from /Mid1, LoKr1/. The phase error distribution density, given there, is:

(8)

where is the coherence at point k. The corresponding distribution function which we also need in equation 7, obtained from /Mid1/ is given by:

(9)

As indicated earlier the phase difference given in equation 5 and showing values between 2 is now mapped into the base interval between . The functional mapping is:

(10)

The distribution density of the -mapped phase error difference can be obtained by letting:

(11)

Using the correct functional mapping in each of the intervals and exploiting the fact the conditional probability density of a functionally mapped variable only consists of a Dirac impulse we have:

(12)

Now we return to the sum in equation 4 and introduce the non-mapped discrete difference by:

(13)

Further introducing the conditional density of conditioned on the fact that the true phase derivative takes on the value we may write:

(14)

In the last equality we have used the fact that the -mapped phase error difference is independent of the true phase derivative . Substituting equation 12 into 14 we obtain:

(15)

This is the conditional distribution density of conditioned on the fact that the true phase derivative takes on the value . This variable is -mapped again to yield (Equ. 4). Now utilizing the same arguments and reasoning as before, we get the final result for the conditional density of the 'mapped' phase derivative, conditioned on the fact that the true phase derivative takes on the value :

(16)

where the short hand expression has been utilized for convenience. This expression is the 2-cutout of the sum of three shifted replicas of the distribution density (cf. Equ. 15):

(17)

Figure 1 demonstrates the generation of for an arbitrary density :

Figure 1: The generation of as a cutout of three superimposed densities

To simplify the further derivation we introduce the bias error of the 'mapped' derivative:

(18)

If we now evaluate the conditional density of this error conditioned on , we can evaluate any stochastic measure of the bias conditioned on any value of the derivative, which we seek, e.g. the conditional mean of the bias. From probability theory we know that:

(19)

Substituting the identity of equ. 19 into equation 16 we obtain the wanted conditional density:

(20)

with given in equation 17. Figure 2 demonstrates the meaning of equation 20 graphically:

Figure 2: The generation of the conditional error distribution

With the help of figure 2 the conditional expectation of the bias error is readily calculated:

(21)

1. As our first case we will consider the interval . For this case we can subdivide the integral into the following two parts:

(22)

2. The second case is given by: . Here the following sequence of operations is valid:

(23)

Since is an even density function symmetric around zero, the corresponding distribution will show the following symmetry:

(24)

From inspecting equations 22 and 23, respectively, we conclude that the conditional mean of the bias error is an odd function with respect to the nominal value :

(25)

1.2 Preliminary Observations

General observations:

Summarizing equations 22-25 we note that:

  • The bias error and true value have opposite signs
  • The bias is an odd function with respect to the true value
  • Thus the estimate of the phase derivative computed from modulo-mapped phases will always be biased towards lower absolute values. The phase slope is always underestimated.

Limiting cases:

a) Maximum Phase Slope

Inspecting equations 22 and 23 we note that the bias error clearly depends on the value of the true phase slope. Knowing that:

(26)

we conclude from equations 23 and 25 that:

(27)

Thus phase slopes of = will always - even for good coherence - be estimated with the maximum possible bias error of -sign()!

b) Ideal Case: Perfect Correlation

In this case we have:

(28)

and using equations 22 and 23 we note that:

(29)
c) Worst Case: No Correlation

In the zero correlation case we get a uniform distribution density and a ramp distribution:

(30)

and from equations 22 and 23 we get the final result:

(31)

so that in this case the estimated phase slope will always be zero, which is quite remarkable.

1.3 The General Case - Numerical Results

From equations 22, 23 we conclude that knowing F0() is completely sufficient for determining the bias error. If we furthermore restrict us to the case of phase slopes between , we can utilize equation 17 and write:

(32)

where in the last equality we have only used the usual symmetry properties (cf. equ. 24). For convenience we will furtheron restrict ourselves to the case of positive slopes so that we can substitute equation 32 into equation 23 and write:

(33)

The distribution density is periodic with respect to 4. This means that we can expand it in a Fourier series:

(34)

Substituting equation 34 into 33 we readily obtain:

(35)

where the Fourier coefficients are given by:

(36)

Since we do not want to solve the integral analytically we approximately calculate dm by FFT-techniques by:

(37)

is the sampled continuous density where:

(38)

The continuous density is, as indicated by equation 6, the continuous correlation:

(39)

The discrete equivalent employing the sampled versions of the individual densities is given by:

(40)

Realizing this discrete convolution as a cyclic convolution we carry over to FFT-techniques by writing:

(41)

The result of equation 41 can be easily obtained in the frequency domain by letting:

(42)

where the Fourier transforms are calulated by::

(43)

Then the rule for approximately evaluating the bias is:

(44)

Finally we obtain the solution for negative phase slopes by (equation 25):

(45)

Equations 44 and 45 provide the final result and form the basic framework for evaluating the bias error depending on the phase slope itself as well as on the form of the densities. These densities depend on the degree of coherence or on the SNR of the interferogram (the quality of the fringes). In the following we will give some quantitative results. We will assume identical distribution densities for two successive points. Figure 3 shows the wrapped phase error density for different coherence values.

Figure 3: Distribution Density of the Phase Error Figure 4: Distribution Density of the unwrapped Phase Slope Error

Figure 4 shows the distribution density of the unwrapped phase slope error for different degrees of coherence. Figure 5 shows the outcoming bias over the true phase slope evaluated for different degrees of coherence.

It is completely obvious that the maximum allowed phase slope that may be estimated with negligible bias strongly depends on the coherence. If the coherence is one, there is no phase slope bias as long as the phase slope is less than . The other extreme is a coherence of 0.1. In this case the slope bias is considerable even for small slopes.
Figure 5: Bias Error over Phase Slope

2.0 Approaches to solve or circumvent the problem

The following chapter will give a short overview about how to avoid or circumvent the problem of biased phase derivative estimates.

  1. Do not apply any filtering or averaging techniques to the phase slope!

Any filter operation to remove the stochastic influences from the phase derivative will produce an estimate which is 'nearer' to the conditional mean, which is not identical with the true phase slope. On the other hand the simple branch/cut methods which do not apply any filtering to the phase slope estimates provide noisy but unbiased phase estimates. If any filtering is to be applied it should be applied to complex data rather than to the phases or phase slopes.

  1. Keep the phase slopes as small as possible by successive flattening techniques!

Since the bias of the slope estimation clearly depends on the slope itself, one method to keep the bias small might be to keep the phase slope small. This can be achieved by successive flattening. The biased phase estimates obtained in the first run are utilized to 'demodulate' the complex interferogram. Then the residual phase slope is estimated again, this time with a smaller bias. The unambiguous phase is generated and used again to demodulate the interferogram. The whole loop is repeated until the bias has been reduced satisfactorily. Such a procedure (as reported in /For1 / clearly yields asymptotically unbiased phase estimates.

  1. Correct the systematic error by subtracting the phase slope bias!

With the results given in equations 44 and 45 it should be possible to further utilize the phase slope estimator based on finite differences of wrapped phases and to eliminate the bias by simply subtracting it. This method would be advantageously applicable to homogeneous scenes with a sufficiently smooth coherence distribution, since then the densities would not have to be calculated and Fourier transformed for any individual pixel. In this case the additional computational burden would be moderate. The bias estimation of equations 44 and 45 would be the key to maintain Linear Least Squares phase unwrapping approaches.

  1. Do not calculate phase slopes by finite differences from modulo mapped phases!

Clearly the best solution to a problem is an approach which prevents the problem from arising. This can be achieved by applying unbiased phase slope estimators. All these estimators share the common property that they operate on complex data rather than on the phases. Either they use real and imaginary part of the interferogram and exploit the argument of a complex correlation kernel (known from Madsen's Correlation Doppler Estimator, as proposed by /Bam1/) or they operate in the power spectral density domain (such as the Local Spectral Mode Estimator, proposed by /KrLo2, LoKr2/). Another approach would be to use a nonlinear estimator in form of an Extended Kalman Filter which does not explicitly differentiate any mapped phases (as proposed in /LoKr1, KrLo1/). Recently a combination of local slope estimation and Kalman filtering techniques has been proposed in /KrLo2, LoKr2/. This combination seems to be the most powerful approach to phase unwrapping, yielding unbiased and nearly perfectly noisefree unwrapped phases down to coherence values of 0.3 without any prefiltering!

Conclusions

The paper has presented the analysis, the derivation and evaluation of the estimation bias if the phase slope is determined form modulo-2 mapped phases. A finite series representation for the phase slope bias has been given, where the coefficients can be easily determined from the FFT spectrum of the distribution density of the wrapped phase error. This kind of distribution density is known for a lot of special cases even if 'multi-look' prefiltering is applied. Thus the technique is widely applicable. The numerical results calculated with the approach are in perfect agreement with the expectation. Finally some hints to circumvent or solve the problem were presented.

Acknowledgment

The work reported in the paper is an integral part of our activities on phase unwrapping in cooperation with DLR and has been funded by the DARA (project „Rapid", grant number 50 EE 9431) which is greatly appreciated.

References

Middleton, D., 1987:
Introduction to Statistical Communication Theory, Peninsula Publishing, Los Altos, pp. 396-410. ( Mid1 )
Loffeld, O., Krämer, R., 1994
'Phase Unwrapping for SAR Interferometry', Proc. IGARSS'94, Pasadena,, pp. 2282-2284, (LoKr1)
Krämer, R., Loffeld, O., 1996
'Phase Unwrapping for SAR Interferometry with Kalman Filters ', Proc. EUSAR'96, Königswinter, pp. 199-202, (KrLo1)
Krämer, R., Loffeld, O., 1996
'Local Slope Estimation in cooperation with Kalman Filtering Techniques', DLR Workshop on Phase Unwrapping, (KrLo2)
Loffeld, O., Krämer, R., 1996
'Local Slope Estimation and Kalman Filtering', PIERS'96 Symposium, Innsbruck, 1996 (LoKr2).
Bamler, R., Davidson, G., 1996
'The mystery of lost fringes in 2D-Least-Squares Phase Unwrapping', Piers'96, Symposium, Innsbruck, 1996. (Bam1)
Fornaro, G., Franceschetti, G. Lanari, R., Rossi,D., Tesauro, M., 1996
'Phase Unwrapping by using the Finite Element Method (FEM)', Piers'96, Symposium, Innsbruck, 1996. (For1)

(Conference Program) (Participants) (Abstracts and Papers)
(Contacts)

loffeld-et-al/loffeld-et-al.html~100644 23423 23424 101370 6220471272 15714 0ustar luziguest
esa FRINGE 96 RSL
home search

Estimating the derivative of modulo-mapped phases

Otmar LoffeldZentrum für Sensorsysteme (ZESS), Universität-GH-Siegen, Paul-Bonatz-Str. 9-11
57068 Siegen, Germany
loffeld@wiener.zess.uni-siegen.de
http://www.zess.uni-siegen.de
Christoph ArndtZentrum für Sensorsysteme (ZESS), Universität-GH-Siegen, Paul-Bonatz-Str. 9-11
57068 Siegen, Germany
arndt@nv.et-inf.uni-siegen.de
http://www.nv.et-inf.uni-siegen.de/pb2/www_pb2
Achim HeinZentrum für Sensorsysteme (ZESS), Universität-GH-Siegen, Paul-Bonatz-Str. 9-11
57068 Siegen, Germany
hein@nv.et-inf.uni-siegen.de
http://www.nv.et-inf.uni-siegen.de/pb2/www_pb2

Abstract

Nearly all known phase unwrapping techniques try to unwrap the mapped phases by a sequence of differentiating, taking the principal value of the discrete derivative and integrating again. This procedure, conceptually appealing as it may appear, however, yields strongly biased phase derivatives and thus strongly biased phase estimates as the paper will show. This bias is a serious drawback of any differentiation of modulo mapped noise contaminated functions. The problem lies in the methodology rather than in the wrap around effect, as no sequence of linear and nonlinear operations may be altered without seriously affecting the result. It can be shown mathematically, that computing the discrete derivative of noisy modulo-2 mapped phase yields estimates of the unambiguous discrete derivative, which are always biased towards lower absolute values. Thus phase slopes are always underestimated. It can be shown further that the bias clearly depends on the phase slope itself as well as on the coherence. The paper will present the theoretical analysis, and give some hints of how to circumvent the problem.
Keywords: Phase unwrapping techniques, phase slope estimation, biased phase estimates

Introduction

The determination of the unambiguous phase from noisy observations of complex angularly modulated signals is an unsolved problem in general, especially if phase and amplitude are mutually uncorrelated or even independent. This is clearly the case for a complex SAR interferogram. In terms of signal theory a SAR interferogram can be considered as a complex, simultaneously amplitude and phase modulated 2D signal with non Gaussian error statistics. Usually the wanted interferometric phase is obtained by a simple tan-1 operation delivering phase values within the principal interval (e.g. -, , depending on the kind of inverse function). These phases contain all the information needed for the generation of digital terrain elevation maps of observed areas but they do not contain that information in an unambiguous way as any absolute phase offset (an integer multiple of 2) is lost. Furthermore they are subject to phase noise coming from the superimposed amplitude noise in real and imaginary part of the InSAR image. The process of resolving these phase ambiguities is usually called phase unwrapping which in terms of signal theory is simply a two dimensional phase demodulation problem. Nearly all classical approaches to phase unwrapping, known from optical interferometry apply a sequence of differentiating, taking the principal value of the discrete derivative and integrating again along specified paths. The paper will show that such a sequence of operations always yields more or less badly biased estimates of the wanted derivative of the unambiguous phase. This especially applies to a combination with Linear Least Squares techniques which are commonly used to reduce stochastic phase errors. The paper will present the analysis, the numerical, and the analytical evaluation of the phase bias depending on the phase slope and on the signal to noise ratio. Finally some hints to circumvent the problem will be given.

1. Theory and Concepts

1.1 1D Stochastic Analysis

Let the observed phase obtained from a tan-1-operation form real and imaginary part of the interferogram, be related to the unambiguous phase by the following mapping:

(1)

where (k) is the true unambiguous phase at time or point k, is the phase error and the bracket indicates the operation of taking the principal value of the argument phase term in a way that:

(2)

As a result of equations 1 and 2 the observed phase always lies within the base interval (-,. This effect, arising anytime when the interferometric phase is computed by a tan-1() operation from quadrature and inphase component of an interferogram, is well known as the wrap around effect.

Nearly all known phase unwrapping techniques with the exception of /LoKr1, LoKr2/ try to unwrap the mapped phases by a sequence of differentiating, taking the principal value of the discrete derivative and integrating again.

The operation of forming the discrete derivative yields:

(3)

In the sequence of operations in equation 3 we have only used the fact that adding any integer multiple of 2 does not change the result of a modulo-2 operation. are the phase errors at points k and k+1, respectively, mapped into the base interval (-,. The stochastic properties of these errors, namely distribution density and second order moments are known. Now again using the same identities in equation 3 as before we can further write:

(4)

where the last equality has exploited that along normal integration paths the true phase variation's modulus is always smaller than . Normal refers to those integration paths which do not cross a cutline. is the true discrete phase derivative.

Equation 4 clearly expresses the error which we commit when forming the discrete derivative from modulo-2 mapped noisy data. If there were no phase error present the result would be totally correct, but since phase errors always occur in normal interferograms, we commit a systematic error when 'differentiating' modulo-2 mapped data. The phase derivative will be - depending on the noise - more or less badly biased and so will be the unwrapped phase if the biased derivative is integrated again.

The further investigation of the bias error will be organized as follows. In a first step we will investigate the stochastic features of the phase error difference in the second term of equation 4. Then we will evaluate, how a modulo operation, which occurs twice, changes the known distribution density of random variable. Let us introduce the non-mapped phase error difference variable by:

(5)

Obviously the numerical values can vary between 2, as any of the terms in the difference can vary between . Assuming that the phase errors of two subsequent phase samples are independent random variables with identical and symmetric distributions the resulting distribution density of the phase difference is the correlation product of the individual phase distribution densities. Thus we have:

(6)

Later on we will need the distribution function rather than the density. This function is simply obtained by:

(7)

Conceptually equation 7 can be solved since the individual terms are known from /Mid1, LoKr1/. The phase error distribution density, given there, is:

(8)

where is the coherence at point k. The corresponding distribution function which we also need in equation 7, obtained from /Mid1/ is given by:

(9)

As indicated earlier the phase difference given in equation 5 and showing values between 2 is now mapped into the base interval between . The functional mapping is:

(10)

The distribution density of the -mapped phase error difference can be obtained by letting:

(11)

Using the correct functional mapping in each of the intervals and exploiting the fact the conditional probability density of a functionally mapped variable only consists of a Dirac impulse we have:

(12)

Now we return to the sum in equation 4 and introduce the non-mapped discrete difference by:

(13)

Further introducing the conditional density of conditioned on the fact that the true phase derivative takes on the value we may write:

(14)

In the last equality we have used the fact that the -mapped phase error difference is independent of the true phase derivative . Substituting equation 12 into 14 we obtain:

(15)

This is the conditional distribution density of conditioned on the fact that the true phase derivative takes on the value . This variable is -mapped again to yield (Equ. 4). Now utilizing the same arguments and reasoning as before, we get the final result for the conditional density of the 'mapped' phase derivative, conditioned on the fact that the true phase derivative takes on the value :

(16)

where the short hand expression has been utilized for convenience. This expression is the 2-cutout of the sum of three shifted replicas of the distribution density (cf. Equ. 15):

(17)

Figure 1 demonstrates the generation of for an arbitrary density :

Figure 1: The generation of as a cutout of three superimposed densities

To simplify the further derivation we introduce the bias error of the 'mapped' derivative:

(18)

If we now evaluate the conditional density of this error conditioned on , we can evaluate any stochastic measure of the bias conditioned on any value of the derivative, which we seek, e.g. the conditional mean of the bias. From probability theory we know that:

(19)

Substituting the identity of equ. 19 into equation 16 we obtain the wanted conditional density:

(20)

with given in equation 17. Figure 2 demonstrates the meaning of equation 20 graphically:

Figure 2: The generation of the conditional error distribution

With the help of figure 2 the conditional expectation of the bias error is readily calculated:

(21)

1. As our first case we will consider the interval . For this case we can subdivide the integral into the following two parts:

(22)

2. The second case is given by: . Here the following sequence of operations is valid:

(23)

Since is an even density function symmetric around zero, the corresponding distribution will show the following symmetry:

(24)

From inspecting equations 22 and 23, respectively, we conclude that the conditional mean of the bias error is an odd function with respect to the nominal value :

(25)

1.2 Preliminary Observations

General observations:

Summarizing equations 22-25 we note that:

  • The bias error and true value have opposite signs
  • The bias is an odd function with respect to the true value
  • Thus the estimate of the phase derivative computed from modulo-mapped phases will always be biased towards lower absolute values. The phase slope is always underestimated.

Limiting cases:

a) Maximum Phase Slope

Inspecting equations 22 and 23 we note that the bias error clearly depends on the value of the true phase slope. Knowing that:

(26)

we conclude from equations 23 and 25 that:

(27)

Thus phase slopes of = will always - even for good coherence - be estimated with the maximum possible bias error of -sign()!

b) Ideal Case: Perfect Correlation

In this case we have:

(28)

and using equations 22 and 23 we note that:

(29)
c) Worst Case: No Correlation

In the zero correlation case we get a uniform distribution density and a ramp distribution:

(30)

and from equations 22 and 23 we get the final result:

(31)

so that in this case the estimated phase slope will always be zero, which is quite remarkable.

1.3 The General Case - Numerical Results

From equations 22, 23 we conclude that knowing F0() is completely sufficient for determining the bias error. If we furthermore restrict us to the case of phase slopes between , we can utilize equation 17 and write:

(32)

where in the last equality we have only used the usual symmetry properties (cf. equ. 24). For convenience we will furtheron restrict ourselves to the case of positive slopes so that we can substitute equation 32 into equation 23 and write:

(33)

The distribution density is periodic with respect to 4. This means that we can expand it in a Fourier series:

(34)

Substituting equation 34 into 33 we readily obtain:

(35)

where the Fourier coefficients are given by:

(36)

Since we do not want to solve the integral analytically we approximately calculate dm by FFT-techniques by:

(37)

is the sampled continuous density where:

(38)

The continuous density is, as indicated by equation 6, the continuous correlation:

(39)

The discrete equivalent employing the sampled versions of the individual densities is given by:

(40)

Realizing this discrete convolution as a cyclic convolution we carry over to FFT-techniques by writing:

(41)

The result of equation 41 can be easily obtained in the frequency domain by letting:

(42)

where the Fourier transforms are calulated by::

(43)

Then the rule for approximately evaluating the bias is:

(44)

Finally we obtain the solution for negative phase slopes by (equation 25):

(45)

Equations 44 and 45 provide the final result and form the basic framework for evaluating the bias error depending on the phase slope itself as well as on the form of the densities. These densities depend on the degree of coherence or on the SNR of the interferogram (the quality of the fringes). In the following we will give some quantitative results. We will assume identical distribution densities for two successive points. Figure 3 shows the wrapped phase error density for different coherence values.

Figure 3: Distribution Density of the Phase Error Figure 4: Distribution Density of the unwrapped Phase Slope Error

Figure 4 shows the distribution density of the unwrapped phase slope error for different degrees of coherence. Figure 5 shows the outcoming bias over the true phase slope evaluated for different degrees of coherence.

It is completely obvious that the maximum allowed phase slope that may be estimated with negligible bias strongly depends on the coherence. If the coherence is one, there is no phase slope bias as long as the phase slope is less than . The other extreme is a coherence of 0.1. In this case the slope bias is considerable even for small slopes.
Figure 5: Bias Error over Phase Slope

2.0 Approaches to solve or circumvent the problem

The following chapter will give a short overview about how to avoid or circumvent the problem of biased phase derivative estimates.

  1. Do not apply any filtering or averaging techniques to the phase slope!

Any filter operation to remove the stochastic influences from the phase derivative will produce an estimate which is 'nearer' to the conditional mean, which is not identical with the true phase slope. On the other hand the simple branch/cut methods which do not apply any filtering to the phase slope estimates provide noisy but unbiased phase estimates. If any filtering is to be applied it should be applied to complex data rather than to the phases or phase slopes.

  1. Keep the phase slopes as small as possible by successive flattening techniques!

Since the bias of the slope estimation clearly depends on the slope itself, one method to keep the bias small might be to keep the phase slope small. This can be achieved by successive flattening. The biased phase estimates obtained in the first run are utilized to 'demodulate' the complex interferogram. Then the residual phase slope is estimated again, this time with a smaller bias. The unambiguous phase is generated and used again to demodulate the interferogram. The whole loop is repeated until the bias has been reduced satisfactorily. Such a procedure (as reported in /For1 / clearly yields asymptotically unbiased phase estimates.

  1. Correct the systematic error by subtracting the phase slope bias!

With the results given in equations 44 and 45 it should be possible to further utilize the phase slope estimator based on finite differences of wrapped phases and to eliminate the bias by simply subtracting it. This method would be advantageously applicable to homogeneous scenes with a sufficiently smooth coherence distribution, since then the densities would not have to be calculated and Fourier transformed for any individual pixel. In this case the additional computational burden would be moderate. The bias estimation of equations 44 and 45 would be the key to maintain Linear Least Squares phase unwrapping approaches.

  1. Do not calculate phase slopes by finite differences from modulo mapped phases!

Clearly the best solution to a problem is an approach which prevents the problem from arising. This can be achieved by applying unbiased phase slope estimators. All these estimators share the common property that they operate on complex data rather than on the phases. Either they use real and imaginary part of the interferogram and exploit the argument of a complex correlation kernel (known from Madsen's Correlation Doppler Estimator, as proposed by /Bam1/) or they operate in the power spectral density domain (such as the Local Spectral Mode Estimator, proposed by /KrLo2, LoKr2/). Another approach would be to use a nonlinear estimator in form of an Extended Kalman Filter which does not explicitly differentiate any mapped phases (as proposed in /LoKr1, KrLo1/). Recently a combination of local slope estimation and Kalman filtering techniques has been proposed in /KrLo2, LoKr2/. This combination seems to be the most powerful approach to phase unwrapping, yielding unbiased and nearly perfectly noisefree unwrapped phases down to coherence values of 0.3 without any prefiltering!

Conclusions

The paper has presented the analysis, the derivation and evaluation of the estimation bias if the phase slope is determined form modulo-2 mapped phases. A finite series representation for the phase slope bias has been given, where the coefficients can be easily determined from the FFT spectrum of the distribution density of the wrapped phase error. This kind of distribution density is known for a lot of special cases even if 'multi-look' prefiltering is applied. Thus the technique is widely applicable. The numerical results calculated with the approach are in perfect agreement with the expectation. Finally some hints to circumvent or solve the problem were presented.

Acknowledgment

The work reported in the paper is an integral part of our activities on phase unwrapping in cooperation with DLR and has been funded by the DARA (project „Rapid", grant number 50 EE 9431) which is greatly appreciated.

References

Middleton, D., 1987:
Introduction to Statistical Communication Theory, Peninsula Publishing, Los Altos, pp. 396-410. ( Mid1 )
Loffeld, O., Krämer, R., 1994
'Phase Unwrapping for SAR Interferometry', Proc. IGARSS'94, Pasadena,, pp. 2282-2284, (LoKr1)
Krämer, R., Loffeld, O., 1996
'Phase Unwrapping for SAR Interferometry with Kalman Filters ', Proc. EUSAR'96, Königswinter, pp. 199-202, (KrLo1)
Krämer, R., Loffeld, O., 1996
'Local Slope Estimation in cooperation with Kalman Filtering Techniques', DLR Workshop on Phase Unwrapping, (KrLo2)
Loffeld, O., Krämer, R., 1996
'Local Slope Estimation and Kalman Filtering', PIERS'96 Symposium, Innsbruck, 1996 (LoKr2).
Bamler, R., Davidson, G., 1996
'The mystery of lost fringes in 2D-Least-Squares Phase Unwrapping', Piers'96, Symposium, Innsbruck, 1996. (Bam1)
Fornaro, G., Franceschetti, G. Lanari, R., Rossi,D., Tesauro, M., 1996
'Phase Unwrapping by using the Finite Element Method (FEM)', Piers'96, Symposium, Innsbruck, 1996. (For1)

(Conference Program) (Participants) (Abstracts and Papers)
(Contacts)

loffeld-et-al/loffeld-et-al.zip100644 23423 23424 365702 6220264007 15363 0ustar luziguestPKVD!Wloffeld-et-al.doc[ Y8NN8iM$9+_qvuӸizь{:K69~۔*95ΝGW993q"~S'q /vsqK!ٍKR#v4=9|yrʹιhʎ5J{N~ԽmSϝ:?=pl& (f3sI#=V|nf\hh͗^lpgjcNy s61qǿPc4b뜩-ね]^= =mM󪻎˼?շ'K(|WPY*|>|U}^qA`?0p>< ^<Go{?b,m uK}dIf^?8o6@p0Ytey}'O'MN:=8|z /K/ i+wx5_|= ^7 | ov; | |?< !~1ǁ~xSO?c?v/2+:;w~-ශ ~?_|kC |;'b ɒ|G?{Jئ dEF`p)pp9P6o9lq= bݻo}|.< T*Pw@vf=q3?uup EF`p)pp9Pڠkہ-u [Ӏn/ڠ@*@ۀYl^v{=^^> 6h{ه~`8.62s#8GcqӀ<^< x 8xxj I@hh@PS]ߠ4 ew^xpiq3zy:>^yLi׵c XGT3 G%悥+s#\L>v9ӫLT(wnx h,ҏ0Zbٕ|LtW2Ty%_ī>n|:5:;[s՝u=$î믔:R5W߮_U#֒~_.m5p 7n׈mƥ zX/z7")QIJDF"WD"HMDH^)E-nQn*"vxL< T=ˮsKEP CǖpP~,CFkb" 4e+ eE4Lۋ,0#bPj~ +%dQ+JzH 01q'YVŪtcd( l08B`VnsVs7h%^`9 aUPN&,נɮQ;A+v@[q}{AUJЌ+?(̜~+} P5b7 -/ZX=&MbV m Q5q(MЂ'U>U$ Q;n/I$bV ie(\O݆ZLXWDwW"+8!N`Kc: ϲ>40<(?x .bZ\h'St4=m-ٓ>s@#I~} sӛAT)d*!9 b>$z+*,Eru`?PJBe;ZGd{JzSe+YE҂B6c3ZX0 ex˓gcEjЖc\y2#xPٓMHʹ;ë8iIVnb,7C BDFKT ɋnJ!Cxn ^^-ۭDFЅ4ǷV3 g!r(RD!(bmLNF\Hi UL6X<neY[5ǎΝ:phٺx-ίtuKElE"P>EzH.Q$P[lBzK- ғZHC 66S8ә%j)F܆vAX9 Rõ(He "AAH(ip\^bI#5%p%'*U %I!Car1^ 9`f C YGى h in,]%z@s@s 4r\mPEf+Ua0SMTXX@6J8 OÁOpP7%pWh(yvp"Tߖ?)t x\„vxѡ.3шKQ#d.%0V˨TLF:5CȉPN\QkݮΦ 1\H,1sCf(*dqS0`*EK)H3"$CDTgHrb#lJq% f^::YH"+c E ҠW)%W9*kr4G!tFAɌOɴ-md(<i&+RFiz6eJUcN`SE'6k$xGԊljmVi >@"&t(tu^b>g 9~oՋER__O?H8uN7wPA$$ʉ3j .Y]<|fxezM a>Pu8)VuLvcCQikMLU4RrQY _,䌜-śYpYWYP= WU[zG~ Uӛ.m2"];mGTd°*JԱ5E i=͝Ē&#BҰit=c`>U]MhnE/Jk6J|̨z<;ut 1M$7dEd;mq5J|r!D7'in.mh˙ą~Hqfs~17n,-:]Vm7ϋ+W Z~ltyRަ(ŲuxS5fΊm3P!ûp|?0<=QljT˥U`BS&RfOmb+EXVB!Lb:=:dhWg8B"TՎS||v7ԇX\~X ^Gau!UeY`4MzTے:<FYt1z3 ;Km $i:4he|b%L B9usͭqJ4?l[FjvxjR&ϫ Hٳ,\Ur1{YgeTZ]e͠s  |OKO#Q(^Jk.SzuȞ趐Etml,.S]=~,B]!:\-.سEf}__@"6ݒ :aij F[Ü kfXRQ]x>P7"RA y`a_'*Ћ[.8gAA"'MGЫ߾]uEVb"t6i^kqFgРϼA%q7DX!3(lDhuij ;ORO[.yG5ນkо eo5rl+fzS J'Xu^[oQ|]Ҵ_=|Ts E-KB@ $@Q;sͻK ` Q멠;v{}oߕ}$}3;{1P8P.g^OtY>Q"=hSO!RWpJ|뎩͖q7*0q=8%" ICZ^=o; h&_j-kިo;,59=ޤQW*RG{Tf*Hx=4TB֓4;]vQӑBv&-dkU34w. |ȝcZک"-0QWM,b` yH %cڏ w%LRjޟnxfne ֫954ݛdw@hСC^:Flx'l:2EتG<0fSl(Y0ү;r{FP %ЙW@ʂTaq|جY,eTbi`BKEtx|HP''wnÙ,9*g!]*2T n*ɐFJ6&%kW4)IYuK:`)G:6O`7V1\TIub ǃdH6?x|Nis;S%ĤA-{eݡqѴcisӶHƥQˊ Ѡ%Gkb4H"0IƞUGȉ VŽ!-猒-̐ WᙼZP2EnEǔ)^鋳sh2ZRyf9?He4T;DMeg]+.k"!/cDuW*QtH x, ܭV_x2=䟗(e,%G „ pME4nj N ?AQb !1p\r*<"cHU^G*y-?m6u VGQꔧͥيjK*Д//ެ B5A}+ $]H°=as7qaz`~ ܙ( w{Q]ݝ|O]"c 1B19)x`HRah]T.J:Wx(S9xTcU +S,J]_ 0RK-܏[j뎶ca~5Ծ/JpGچWQ;]^V,5҆ ܊: Ԫ'KD6o}D4Qoxȃ]$R\,ޖQftĪ%u$'rmCX5wS|+<&)/?RGo (elca|H7"a]|=0qB6LE-3]D;pT[yyIꚍ dJgFh6OLX^,WShS3&%OH,JP*!Ŕ'mR(Jۥ@\3mW^!t)K"V8rȑ}9zt)z ߩ2iPw D0V7z\@OH`󕊏!nؤ ۡGNz:uR$(L5q^Q|tW> % l[q֤{5Zol[>حJqlw4)#  S $o^;uHkhHuմ*p( `lX u9F bq)"8]p'B:5S?@]AQ66S^AXZ$LqAWt6xxª)L+U,pVG7x5`NwYrGgZ>RPw`4L*ǖ-!;Κ$MkjwyEooxrՖHU5;mZ61_C8[*M Ҷn- j p^_Ril1eR!NR)L$uDU nxTU0?UAz4;֬%(rj\gcnVtEń1<)Ҿ.=eӌL1IAnycmxR_BB!0H1c*ʈ/VaTa:vҠ*CA mamb}DEk*cTwdH  " u.6ٽ,!$y[dSQ\j'KV&z YtR6 Zi^6.Qv#I*jPz W+!F' #RI*BMNKƈDBcS7Ɨ3U[A=L53˫!\l$j[*(X#NXoQb_NFΈl@ans ׬S47=4uz4wyc5!t"2fq.)x\zYXv; q@&|^tL,-&'JjvJCfֿ5n4WM} pjs^̪qֺB{B[\L[0R(:Zji[0n&:ϱH@2qw29A!Q/lcGg7W^V2Kٿ:|yj6,W,W3]t55YpKA8wѽ VN0qʞ덗>3wKVɰ2dA0e)tzrH,:}k`Fe+>|aq Æ QFq`̙wuYn:ذa$ {wށ_ K}=;HnHo\OBWՑ8,t:tB{;!] ~yJ!uB3!3sB]Iӛ9v2lmGl1} B0M< vGdGaBE>Y)u6DG_b6Мoĺ'4Qh`EÖef,cd!rH``RV in1z "4 WBQG ,t,mxdIp[86^Zh*;0Y`)K΀G䆙ۧnd1y Ѱ9aoS[>${~a>0|KÜ4Wvvdq I"Yh4D83ja}4hv H; 2J\>|81 ??vyge]`ȑ뮻n I aw=sOk`"(..ѣG>%%%0fI2(// ;v,TVVR[r„  a&N&M8&O 5550e:u*B]]ôi`0c hjjfjΚ5 ZZZgφV3g̝;͛t|0|8CC; ,X~8qvx^/ttt@gg'|>pGBWW C$: (b1 ==="-^Xx1,Y>h8ccvK.e˖aŊp 'ʕ+ON: N>d8SSOUViWO?8 Xf ]D >F?ʿQlؒp y>rfQ ('Gl1i .شb>9< bXZAU!;`>Ov0ܜt)A=0S#XҥHi\erRRd4y8$`&a? =2ØE0d.=3+[@:M{o(bz %v^ǐ#XN!CC_leĂ4D\r_hh"Cq@w^`q VҪ2y^9B^RfPMa#xYK^#PnqEG^yFSR grG.(3KIym :[O٘Tż|?\`$>0 e?L ;3vFfS/6Eөͳbv.·] ]a+X3{v -[?O4Y2W9&;I@)0Cs h5fsŻm#_gtP0GEEaaO7,]?XXl)@>cYgRRr2zWRYݷղ~Z\6A6Wn?m~"loK̏'|ts$rsej.sK޳kٲuߍxevN[42 G3$ Kx#D}OSd@fiYyc1"+p> r0tq|~rY20 TF ooVH܇"mPmvy%iʢg3ݚ~&~R: Ҿ=v5!ٙC=%G_\{F=)}O.rzHf ʪuIeF2)IIf>C-kIIޘ|0oQ8LVN46́"+Ǘɴ(8Ҿ"ͦtyU&$ef  3Q8sw m$ ea ı}jCOD*,cPX穲_:Vf@ChVo;}V_Bu|ӧv N% O!@F!`u޹ O>b/]98 rMHoLS^ėLgԠF /ePu.ꨩi8/^ `Rq`R '@Q @Ju|u-V|UXSHJV39JoRUt$:ʊqf_ӷBkW'D:V8ld^ÜRlsJdHdo\c-ז&lK3N*h%x@/5ةC$_d+} 'Nz5 \*ɊGYP\K/X!y(kVe :/ټqQKlr14mo@O|8:/BH*5{!+T!Z@+#d} <1R2C!|ؾGM7< %a'|ASED 5ȅCS^#䝨@kR9D={[7p8TTcG(p4~ᤋ :N38t] 8N5ňJzJz9+WNN5YWߑO 5;ߍ~:cu(f$J/Q&S@!j ; \F:.kq:0\n!v@ă<0N) l`3 ~vp >s|JA% W3ZA 8t

 7Au8 ځc@:FjFp3 <&|x7`h%;d떭RY6Îcj~ Vp z =ǟ NֻDٝ$|Z$6xIGV5{6S' .`&x0< GO68p=xL߂*|ւuT7/4p8} 0wY`6g<<( [HV`p8tg`_OWJY^W~foWPuhbʸ c&r9^Gw kچ]zGÞ@68  \&4tNA6)EA(x <g<<$B0bp:G#`":/@) g (t8` :bt<NCA..A> Wk` wJN+6߸^Je߯Yk'cZX]V<@?6NEܷ #>|'7 &W qg~q96ReWz`M?:g P 0׀ku`X_:wA:^GG PJ2| V:TDp;<3  ! 7`%(߂U`56`\eE%TJg1w.zzllQkT}B{+Nڳ xשׂl}UnYmJyUYWoSujH:99+vk?bsv]YoV4J=6_xhx% \-n_]Pz,Ljcq%;g]A^N!Pn%0/eUyװP[ gYa9A-~cMOG}[xZSWq#{첿~okwGg´(a 9a5?t`^\]2hϺŸ́.~v5D9_a~e_.`wvR}\m[n17,g:fi},5]>QҼ**9?֕31%93oF~mޞ_x<뫠>NU5+3,(S}u7@Uaֱ]z9y#ǎȺDapVq8MRB*dt,%۟>t.S9GGl!;A-N]` gyby>A~gqA"k-_Z7;}/;͝үuWNfAM-g-b3e@}AlA-i:~z_sG떽:X쟕7[mݒ >8|SnѼwG;ۺ 3`sA-j}~oүus軯ZsY栟rl~n56O7[b_Ckݲ{ ݗ?;ʿ-)UcTnY]MO|kݒ$jkݢo cu*Z7?i]3@?cN_~sǜOo殼@}cSc~򄁾fm=NlKhKyThow;spSkݴN?)nƶ'=sЯm@?*<)@}0G;?ZY8ZQc΋s~4ߩiSӺm>`'׺#4Џ?ЩJ_Z,Plݴ~[qW;ۺ/}o{[V}-)nqQף?)Z]?7h;OlaK3ЏSu3ƥ軯xx8Z7w}ONX@}l59M9@uSdkݲա9lڟduEcLo_usw<E r=#([b1? 2VTn_d]۟Ӻ%gQ~losG$}|R*T>t Op?}DHpu.x{/HZx]QEX+DnQɴ:"cj}?)ī,ݷ?Ѻi-ujͼs} O T_aN~'K/sG+,ݴZ-OsI_떝w_}ڿEfGkݒCwO(Ͳt7U?/nwD*Zף#RuEenCI)ZMY >v'wDnnŦ(Bպ=P #Z7m\sV;Z[~_eGMQ8?=guR;֡{vq_?Z$|[Rګ3ϑ(ӈ_8!Wފ5ϛ8W"뗴~_.9gd}j#som~@K!I?bP2{ jwlo>l ZLdb cH?gV k6ڰ֫Z{QGˆ5ꙆzT o겿~]JhuYcQ2B.+a %z5?Թz;uu܆z5L}@et1*xBrc5,LF(ۉԽa7r^?7R~?ި{i<{`s.&m/kU0:S}>Mo{smIbO1?ϋR^?dyz+Z0m3e [)e)H ~[)Z`>{~$? yao&ys#S}//8_/O7>MfOp,:nnmvp]npus&C9STQ&="oQ0<O3= 3Ӣ|.ע Kσ%@W`>x .j8_7[C2]o!>'|?Zi)sX_# X)j]ow=X ։-֏~ fEZb?c{~~s}\kkm}7ϋkq}5S}5L9WgG_z~͏?+g}NlD%j϶kgfOg{^yi;`vOga>4S}UL_ه'GfT5?uplq>^?tS}+=Ty_gSs?ic?cDܵ>sT?{7?Oci]pN?ޣ)S-P 5տ~7Կk?gtj|99mSO|}7~7? LlsοAS}R_5S^?[bz7?jLWl'ykgZ ;5^ο2߰c?l`cLl\&j "}ޗHQ\mpQGi`?첨""$HĜa A#1$n45ǗsM2!f?oW2ڟcCШz.5' +y`z8hOA35Q?QOO'1b?=EϻGg_8GӳQjcj1󏞧g)an zDT$̿a ovy0t b=?s ٘1zot5a;²?$a ώC$F_{W˾D{$jb濁F$x)([ 25?/ۡw>忰Q ˘Oc1y;aM_YEr֓b8 ߜ0oP=$q𿍉wD}?I3}NJބ*~&`w( s 6-^8-Y?_O>ߟ0'b= 3gc >U;kIZ%TOs{3z˾Dj$J{iɾ1B /&g`gJ@ $O'WY9_0h3wFt) Wd97fOq&ީ*,z\1_~%Q=_K؟0;o$x/n`Qw'?+a-EAg&B87Is d?Rmz'fߌe$c $L_3oE!+|??nDz)Q%̿&̿̿_@kSb=~Dրw %߇eW${R܃D$=aߝ?__pT`(C 5^}[Z/b&alJ|/$ %?:a?̎CObٟ(}-'I49 OI qY,zOnߑ0Mo3_@߿,_[{ D%̿[F{}OI8~?8 ڀCA~a$v J8s}˾AL%o7J?p.׷qH%I[_Hbˢ(τR%WIؿ3aߞp)f(_H)ۊ8ߊe$Q ` %);_'N=^oZz߈yXCC%ߐ^ 3GHK ?UmJ=~o'a,}$I=~0 ~eQQ&˝$V7%"^KQi@$a1󟾥K81M&̿ӓ}[1ow}H8N_X#~!Q_KO~yA},;IX%^?U6ڿYOfG;5J~x_I DOC[pǿ.,{D' ?͐6(7vPq+q@AUQ5j+:u0Vi![E5PGu궨ۡnAu$L|s|4NcPMԱPǣN@: ugBuWPwGQ;QިSQNG݇gC BFzLQgJߕķQa{m5}pT1}ELT>+};e⻝PF.@ a}xT}m*}o{5T}Ea}Coi;)FRlTYt |QKwچ=oɹ@$7fa9:pu&ꁨ>Nsۭ!5@ߟ1& ]p5Xdd>dؕ#e`cMf Cdw/H^4,CCm4)MWORA؆Lqcs |ua+xOd{\6L2ѭ6,YMr`i[V0U?Z|s< 62b#eΪFaeŔxeaҀf4W2ea"~R Zmď뀍CV"`x+B,RK}׷bGQlWEO8֡؀gq^t^+I)^Ln=A^)Giw_nx~ FR9 g:O'+!>4p )]VmgoR4q@R)@iƧG +Bh7gne$ ^eG(^C>S`% l&FAau~Uw% a(0'AiFaB/M^"a9\evib}{0VOs|o. x}8s@NB9|n0PuE]Zei|9ŃknT[ds`*y+6<^a`c/H Iy<7zd@ypjYAKY|6 )ŸTZ?+SJ9nۛPh [4h8 I*eB;2_ǼB9l&kCAQj[+6feqW!xPuo92,:PČ+/S7GU{!&!W7WY^|5#8c@gR0`xi(ø֣͛nF$=-V{xIPw5,mqXJ|  KW??*5+-0 ~P-0 ^# tD9O!!s"9#^dR>SjЄLOg;KS, a½gp.Y:biBhxd9e$jj7nHxS1A]4!T{Ă|Ia`{ m3[M7ڼa~}wk~vmΦӛ>lJ,M)Zs#Qy_N}a܋'c/&Ux]l5vJJpj?TĻ(ΖcT}S{o_?hB>J~4RJIVc/)Ԁ@Bȯ"dRՄOS- X_P=9<,]+_ɁmKс0 ńRO[vKyw޹ٞ]Rx t%<@iR9Р݄ψJg; ?LηD  f%U!o?p)YoW3{? &JꆐϠiFP SD (٦ʵL Vd g(OiIϐց糙z0im Z@p(,_:M0h|kK!^\I*x > wA|O?y|Ǵi{M1}rG'k[S.9E{L. {_# fhS,, "\ KAFBa{1[gyӮzKnfFr %oX~ Kjcl)p:ʽ`-Vs^f{+13e1ʪ=EQߡؼxD9>9h};tܚ2ypM$;~TPOqswU ;8|X Wes)}5:<{Y[*nry$]›+`I5 >Gs =Sq/'Bkrd'ϝ &K*|K Obpv:dߗ02]igsԔmٲIG@YPRf߲%82O b]jw-4G@Hm:>>#JpT vG{dy>A) cyR9jJUil٤#A7G@K=m~˖߀q368N4O\'7G@2E٫#\PnF@?]߀~; C%߀.|>fU`6Z <bKP)դ9ڲe?PRf߲%8㫌˚#_ʎvvv ]QfY6I>Lwq4۟ĺX|#fv!]ְ?_X#߀.yʪ6춀WUї+ܻ9jJ5il٤#oescl r\?>˵˜]p 7o@Ӈhn㝽asٳ6 , kwOWܔV {9PtEӝG18+\TU!_*ccUX9Hn嶸<}׉Ygy$V6o/Lgy)p3qU g9Ys#`g|s?4[}}}y+s{}ޕ *JjsG+Yys=~^x_z饗_~W^yW_{_7x7z뭷~wyw7l{?y7nUIS>Ʋ r+QB]z6jsPE=QG Q/b|1#3z ?@^#FO]7W^ԫQz 굨ס 稿@% רף/ AC=PoB#ͨ ϨꭨގzꝨwu-PF;jރz/?PC LC0#>8O>igPE}u=/˨]Q@}-ԷQA}u{A}u# .1`D3hO5gy .Y0ߜ<أO>阓9g{s1'/0Oأ] T 0g {=-bIO 2)v*< 5J g21ɉW rn3o`Y(|2wMiJSI},ޏԺ>ZbXm˭cu~,ޏ7)M2d`_P/S]|uLxpma 3nӹ*.O; (Zn_ᙃW We,~hΏ[*Wt/E|Ӱ^⼪o-xؠӂ -- VS)MiJS>v2SQ6B7k`kW% gҹs}NdΪ-U̓涬.g,>E dȍ M|φK㗮,v׀Kչ^KeK*ۭL\p9v>އY6i\?(?݃ MWNZo ?R[V~ܠfhr;uY- 2+g֓ucu=ݍz= <"P>e LSWo eԑv)4̱z M*sVv];> v<0,V{UXkOZrv ϙLQj\sI*Ag&\_~!2ABnE4capiKk ٖ lۏ5;}x3:>c BxxY>FsǼx, 3 5^L#?2tx;NE+l,K Ez+Fe?ly0't-Zצ~Ә<'XBs_9~ 3P^[S<"Y=bA1bI;s)j`[kńm Z>X,M,Í|\,gð^,#?^tA$M<&INN")O&^-q !9Ёi[4|e[~Dgb߃$B}ot38e~ܕm+ ܶR|ՀȱxxQx<˄wC#bAYǙ\t%?5gUp,?]1!(JkSdUfړ"oPrb"yQ9:>LS:`uUcG[y_iG.`6^n0OմhlxTs P'~xpŹ؞9W_Ϭl;*q">ږr _$K`C!: S,w_7Iayc&gPjtצMS R o yv4߽;wNXSӁ'CI50\i'#;F8&d(3@wI7k$ 8^ίVW X!#u=+Dx2:RV仉O&"1Yje,1o?] W DkAdS.ĐQO&TTH&ZXQ9?z"\&2 Oǁ Rh꽠 s_CygQ~m!J?2؃Bi!i[ iR l6V[DE>kdZ9Ab*^ t4,Q%K!!/ ЊQr=<aѥ[RɌ0S_}!ⓖ=os+-fK!wQg&A36f/dTsi3lS z6xaه %sL5_3vD9rF ] ܢ\AD%0ewHY~9;iȐWGL{-7 Y,9Q!??vU\vg2G:#Qy~r*x 𠐒Bםh=T^w~#Px $ӋSS~7g3hI ?Tr*V3qYD^aqoA<⒯p Za:,/,:ik^U op_kt> 8$O[ ӘqÃ^4;=wGa\@Ȝ ][sΜ _koiVȧk<]!XӺ_)`@es·$|˴@7W@Ι3$JfYP!|kLt.Ӂ=fL%{ LH[0./yw&1:R@(x>8k?' +XcsW0m.2M\Դ-ſHTپ;o$GgBl t=b^Mh8}VOV1WѮÿB]?M=OfL6 =_1wYp&M (o/@KK{F-`:sN/K_N딾1MkSm֩8n mi=ZۘN 2J̵,_bR~Yf=uSSO:ua^*Nro2 l=s'|v=SkS[z*ˬFX_Z*Nstꯘ 5Sdwډz0ZNNݏӯE[tN]wN+SfN}iGթܵރשXԩoGԩ7:n;6ޭN[1ZS`Yڶm_,Z4Z/l[q_חS+S9:u!^%)~2v־`t @p~Iog:Kw/%?c{?{]=hj~n滧C̹2n QK?#Jya%9ær+fߞ}i?<5"C_8{kd';F?[7K ~Îw¡he!5@)}]$#_o-GQ̿i][uY#4C0ā hxl.К>2ABbmT%Z Rp~1%"2߇GCa;Tmi_㥴e5>`'e xM9-*'.#$ _Oo]y=)%Cd|r ?1Le|E|_&#qWߑߑXa*O c c`8bE>^%,|[@/QI&~Yj?U0!T$ۊR WIW|s|+$C9'8_ CygOB<~L?Zf jGW+?@_¹dA-|_@9(?R/Ǫ$Yp'%嫥C~{?34~ V ͘:H0r|Bs9怐5r`b~*~CFv,IM9B$eY1eTXr~2U`9& 2X~[r#Q+?y}bXve5TI(ӑ$3p>JM;^JGkur/K9XOHW_|Sg,R_VBS/~4g B/#QL?sэ@$ /3۷'5B/69\NHesqW|K!~0B)(3寒 hW*?UKW*?U9>V$ ^bRdPӟU?B/%.PJw] ;"/%)E^J"lx"/5>"W<F7꫋T"ugcj-mUH.@*1;U7fa7_k|$B?~uc"YW?I *_gpd?{OY@_O+Pwb)5&^!]c~w=mLvOyo#SEcT{0ZLM@u*锊o};I# ɬ8]މ`hK2LN79t߮YrO75*I)?!6yYzSG9!e41 <~B4N,3j;ZBzA'鳀!yqB^8bT+ƧZ 4snR~LZb_ HU*,#5`ڼDAe/SR!~lC TsA]J0b%\]9pit.P p)tOTb34 b3yϙ߭i{[:3}гZl5JW| xӚV،:\B<qKTpBB=H'Cv26, 61j*q-y?DvmSSYc z?5!ϵ,<΅בN%ã`c[5(CLndIQ!sf؃kLϞAO:f֣WH^pk4:1ᗱԽ:W̧{͓;!{{Fɍ*5# vo=~K &bTxkM.!K~|fsH$|T4:__o9!8IsE!Ga'ac~x"~Å?p }KK/|i`wN!Zhᢅ_k=Ne/u#;Cn9(Xc+ `ESD% mA@ 0 a;>=zܗɟG B'Q;!Ȩ6uhR~?P鷌B -~E\N8??l9Ͽ:]/ , xc'21llL:1qY|z}.΂-'6  NMu4B -$Fԋ d 6JOAְ5!2t3K0c>Î9E!oF ?{О=g;>"ѷ읜;u2epmC۶6,F-=xujs_^Wo>'$KROh$ZX\|. $o ,{0ذ !x&h<.'<{<#?bOd0'8Ha3|82JmwޓVࡇQ{ ﬇wx+w 2d3@3v3F28xky U`"K|T6"mדͻn]`q/-k;ڮZDĥ,&=Cbp ?Zvg 97xF߭ Kc:Z]v FrUrQˈ"˅DDCxx&.OVx7:xxݱ@"*S`'{K"NroEY~L<dZ浶g8˴M>fxK|9ˑOV0otd]XF.Wua%~ߊvGv+]^k]6ri:7!ϻw9кW4=(,g\K6k+e`7itCR0P̛ËN~"6_61>%#e‹U$t}Ui`I}G5=ywllazt;겅Ql ^}ٺ%[ߧfjNw[lTi̪FNe24%wuu%JťLCJG)Ǎ12mɍn4q$;[VTQU2i-AV\ C\[vu>\iZ (@~`U9ݹ4=乳;|;?;3;?zß[QG.R{GOּ-GISxQdgE$e;ajTZZVVVQU.5KMW^Oof b.?KKu -mKPUs+ _sP%Kt ԛy4[diC\)_C[!oG +x0ߍE yY "!Dޅ|5nk Z#EM됯dexGȿA䛐'oFoAȓ!~ېoGA;w!ߍ|ȇ?|ȿ"Y 酜ȟB~ G"?gE,cȏ# '!8!?䧑|K_F I?F><,9r>{hE)f8B4˸>G?rGG$~'=~3';8zDof07'W, |~VGwXJĊk*ڠ4rc}]DWc`,Xp2jXREyR q4 bWX\~]X*+^d p2l,Bcիe,{ƺEj,Y32cj,~eWX:+XTYUpYHFYIR9quVHdw{qHdwqYO^YRæ;;a]f:aX\0M\0갮7ya`O 0'0':a(j^j0B0)BBMa`W]6:vvX0apW劄Na"h]~"P݇9!`xlBr]*ڟ[_ۏ3d>rn7+?oV6xs1ڏTϛ*X1-̦9ٔN}O҉lIhĵ#;<4LތȒ>IȦm4NdU!' NdU)ՊʯL+ʿ+^H˯zLioW<xN3&;|`߰og" :o]]yjG;BtI߀Ϥo"WPe j3w[e Ԯ0&zn[A&)K_*1z ۏԊ"cY#]VW7n^z𜢥Vd/?.U[/J\SVd"F=ȮVX$I*QL<Α=^/+oW~ޭf$J ;-4Jޮ|aXe2P!W_'BC͟V J, ~ >=MOM8q4L Dz|f SkfZen1Dm ѰxCio_~:N`3o/JC!!C*7@k3.%Bd2ד(I*\~zoZwX?uk~DIoKddDBmu LP?b)oJ 1m,5-* c[^eBDP/Xe5*茗ze o=h d]~L@7_h?aA-VxSx ˷UW[w_>/?W6A ~}<3-eumlTi϶&Dơx7i oq(Ϸoq( jA~}♰O|Qg[;I$e^?/lar&ga90N` ؒ?1/^8@`N,L-`3=ay#^l\{O#~| CP5fi>uJa|)# X0o` ?mZ:)*>O/Dy~ZkJ|B-ZbdV$Q;E`T]kA)9aD4: d'r*RR@cq)tLcdV(`f (:A/l-\PW`$U٤T׾M(Ť yAP02LXFBn\EʀoQGn%;o*el =h i?$t X4C!ڕ 8l odQ ;3G^͟<6ZVٓ1^9T}9ÙԒ|L5l/7IS jt:A<f͙8vܣWkJu`ӯс`q!>Pځ*66 E>?"J9r>X2@oZ=9Vȟe9h\jPX< ̏m |~&P & (}ɊFSA<,)F_dƹ)x -@;>^s6)KXaFF=B1\wB%j쓎6U`QɅu0L :E.D m<_gIYa[nE1C4IC@(eJz0{_ pkuf@D4nV;p&(Ù(<6f13qG+0';=i|ޖrJ$ĥq\l̹03i[CpαSf2xVdQ?6xB ϱ6Xmx"Ň"'f{Tλ?5ȉ`_JcSE>?-)ɭeOcb~>l`۫9qz;1/OEi>đ6}h)՘NckťA`T{e^\ߞD[7C5S>5cnoλ4 dq) Xg l |2Tӌ츕P9<ѹr=Ei959(E Ir\jdi|#4 Dw2t zCi7(t:BKO F+Gfd\ߪ,!jAԵSkZoØF]?AShtSBy|gp۽۪eZYvWhu DžV:vֳjx^\%>Jy>ų(k?LGMVCf^EV0eYEBNdt,ԉhu{YǩP'?N?_7i>kDԉl5Nd=(d7IhMڙgDkwUկ ^^7ozC)_4Vi(_0VTJfQDhTZ )JXϟ"%}߃ NׄЫos0l`쿾xoXxq{F_z%?ktfħ0 txV3>H[N0}\d;8Q#ͧ7~Ob8ΎlWd݇<徣JO{" ĒdFw),HȖơZ?A Hں0".&ag1OH lJs)D8o{g!6^1~x}'b &_oION6KL`ԗ/EXNom,yn\V)!? < /-x,?)꼵cO x:[Qk^^t=:ڛ\8%VCb+B#q(%@!!|4|@ BӒ/V{3w3+)rٙ۷ q&v;l1QZr"wUy8폿޽e4zEcp)\0Wt3^lB/ִm.'Lfɛ?G+Xs@i;r4]8-}3c'la.s9GM>g%5_RuQRP(ʹET/EhS((SFԲ%u%%%߈:2(QLf8(SFԲ%u%%%߈:;JjܨFF}F}F)#J%lDI-QR7QR7QRQRQRQR(F#J%u؈ψ:eDI-QR7QRX2QRG'S[৥/[IǨlʫ\EYX9ѡ *{>ʠeRYVF:_>C5/ǨmnxAر\EYWFA܋T[ep5 Zxփ[ypA/yAV Էw/h(w,6N`E^,ۥc6ND7h6QLA˛v%$ d'!;IىjN@v4$ d'!;IىjN@v4$ d'z a; Z! sw3qV~zg!`; d'!;IىjN@v40WSW&lԉc3n3pL[6o.[y-Wz[7.޼3=̘ ~67Ny 6',v|.PX-zUr㙷]vzƊ޼(߳=g=M2ziByr|AYi{PnF@y^[P&QQބfꑞߏ?Ay+Pގ;QnC9.wOQnG3܁0(|.Q>a9GQc(|(DKO|^Ϡ|P>Ճ C<P(k 6e2DUd T$_yi:jCZ<*h3$f!XGf>XpyZRL[gzҁ U!$U2;>'ZWE/0*E(ԠM'vZ:s~]:%5xP?ej°̡c0_zr1UlQխj[@S]Oɂ#kt5upLڇBRjc%P+TsRŒ1/P`ӧ@ҹ: PR0k|{ՄԲZP9'Gw(Ƣ^CJ} X?8,q9`<&5 Tol^' *`c}\K< K*fF"Tf9A{J*J<ـῆ^&xTlsHc\ \ ]w| %ͲLJ/A j c ]zڎ\. 8{n/8V+t~74RMAp:"l}a?bf)n}b[xʞ ^er54)h?PAc #D:`W\'T ޣqE+>^a/5 g<ί\Z #֓&rO?/Mjv773|X8xftH/P GDs9 {\zN Kߩ9$lZ>%k$_ v<%[3!w!61Ȼ;}m&g >}5s|<<9͓{K3m ɾW5-0S xNX^͙LYL)dRÜf0LL~yM Xa䳒}[A>| 5k<C6EUm=yT4)9Ʒjz~~zj1fl_dZk~xi;=b%wqJ8GtYdΉ]+Lk6HӃ=%(frgqu4)踡fs_x>=UӘ2jsX߆n 9 l|'fť1+l.pXk3&wE&m䮰tzt WZ9xec~BZszŘNWMbÿ-·{[yOMmXb-G(c!aW2a=$8MWH}Ϝ?qDZ/;ߓ#xZY7wC-7!zWr)~g:[;W/m]aazՖV#ڒnձC6rUs?D|6y)CÜ ݔr}!|)_Z@n>;l||o~|tpLhyrfW:ZY/l>J6S^~fwQ(߃|&oy*,㷬ۦ9_lx}P_뚈քIΟZMqZDI!)̏Ϗ-_;Sqw);yoI9$O -\*)민|HnZsU?ϗw4_|;?G?_~Вܵ;{?vSN{)7/?>͛o?-oF|VK:=|f' ? fUTi~iSO-3m>I{y܃H-ϟ)w΋ϷO6$OBXSaUaO!au7?m~oϼvSr gXN0~Jav5?:@noW/JҮ?xGws;C /z\ Preor_fxi_y5UqpJBkO" ǂb#$}΍Kf%Tk:tM5\/0:J:G׌,NDKi?ͶN3YG\8+jlT&@jl!²eCa(FظWg[oL8d߅B's|?EOed_sz:w{j:毗{}NOg|tƷI;tb~t>*w%>s?=/5Oq9-2.MM~_s_W>L voLMQ6.폰wqD+/ێ=7km fS,%X"+o׉ǟ٭*0lDڱq1mלnӖmŪL-۶͠R2̀[dEQxTl/>rW9<>ۻ^]6 _@1GJiU<9y]N^Կ|dr}}|5g~w-矘gRF*\er\X?ihj; nh+kbu9Wzt:듧E3]FvV4ߓG{VL 4`>EEfWg}X=3OYXV"96Wk^~<=Ҩ|Euz Rnih~wIɷ1E\ T.ҜӔuu 襓̖#%r_4s6KE)?'U0G4_i@#1njE/}}DY\bSQnTqug`q؜B):y|bL{SxM.?{WEsLd&LI axDT%IqEY B "kԬ"0뢂+J ^(x!j쫪>' mL˫)eqSD GJSB&*yoG܊M'ܨ%)l =*=}=p^LPhiXQI=n3\J UCvR~Q*HYm!&_d||HPE.o*f_ՆBQe NP1 TX9c:ue AY '2"XKYX(7cAC\#4IH"h[K[U.qk,ֵ~xA[ʻ Š4oww#'~?zKUþ Ӈnfpww_.r&sKnGBw.׎l/KLnTXWv\<63ya)r'(wg4=Rҏ̦$Km~)1g6'O-})wor$2nlw]D"ʌZ}j%xߧV35WDO+ܞ╪Pf߭DY]Ɯȃ36c䅹FfB>.]wF^~^v1U6U]Fec*߃ėֽL\ִvj%q~;;0Bmn< M(ŭvtx%UĘh=F3سF懘TZ " o\ɩMThҘަ&A&3{|Z}6ZomfqJL[D*^h0_#|"Gd> P+QZ. P+uSpQ-h;9S 7fLt cc䫲9OtgΧZgLQTyiT>JkwOcǼ㌩|kTUQ TU4tf035,/Q^7?D=Gzke[]ux#fX_wu3%i-5r{|=|>KFV_4ϒy&_5pqZi1l2eEU41rW9oej|ʗ܍1>iH5Ds(l'~Si*Csgxyf!W؋5EjHl9c%3:f%ϙF=cd^CISny)2US2k(u(.erV#o] cw66^y7Ww(T:)Ӣ2+U5z1&oO|TR1 4yHKEV]*L1Ƶ4;/|3NJzr?<322lRrx/HfǢ%r5.f;)'jE~LbLyJ$!BH'Ƽld΍+ cI*3]32RQN~_=SyշI6sXn# `"k8YV:U57Z3,wO xM$zLoњ?Wc,ÄTl#˽,gx*'?LG[2ETO;i\! A+ Ѣ| ֫L3 *p22VKgFpg93aݣ̓~ޫ\{AΫ +աyA?]ѼҬnW,;:?Bfgp ; y GюsQIPPQG92R֕_)6MY[5ٔMYM~ԥL.e~Rwec͇|K^Dkݙ;hQ|Kn'p3RG,~ $ד6:j{Ȫ&3G e*.Id6.cT?JF2yW>^q(g'fJqpGfp,F _PZ[1pĉ~q>Q{n)th r2lvoku A:㬏kkX,ߙ2- %C L01=M&Aa&=Oj?v])s+X >ym=Ga6[4h>e49_a4y8x?RQU2\ 3@1@ޗl,Rǹ.pRF+2e.uܚn,'eZu(EDΚ_^Ϙ#<1k#:_[EW )jhfM]3fy<4\>kL/c:"z_(jjR}?.)pqf[::k{\Mc o6˓.F[ws h;]C1w+^ eo d̗Bd@!3ۜĮNz&_/NuPǘGֆcjZ W(LThSTyNig$kY~ 9u@8u@xx$&)yA%,E T2E׊v_I5FJenI|p¼9-c=n8BN6S>,`)0\-vgގeUy9SwM)H4i밪K5E2QehIIe6'pz,Vڛ{\D lmBrNnwuN SE+v%0^&߲~KP2 K_p\bq[ Oÿ`ih`8dӏSil]]篍y Ζ Ĭ? 3rЃqT fcQ.ŽBm"F)5B.j;GdަW*o#+iW,-<24Z8Fq&DW:j戴Zu9Ka[dJ+TcS|R<[w%:BB\-j H58Մyygldgyj9M 6[Oі6 $V}͞kw8\F2 9 ̔LaLtQFyBhѢ $\umԥ vE|˘f )ޑj%yr"фk-UcUGIʹ8T skV^2.H_7uAuPzh~};x{맢1c?M 6Oe!ŏA0駩׽[S5sk4:1jsTK-N-U jtiHm|6U]Ƣo9%Ԃ>P zSBhXNᙻ,UtJ(7EVrJ(nt6O'6I>fB*N O;.tR 6kՌ   QQѮVtu^<ܞljAE<_j3o\E4oT$X@Kʩ`IK nSۺ8=!\D/( ?6 Ia߷.8E>>M18aL8-@băFd/< D<]ʼvYn?S7Gq~ _ǀ>N?j%ďm?O8pt`3g8P0BB1G/U/\ p5qkkuyAIɀ)bTui@)`:LI7C؟ (,ݤ[k&#ZPƙoI 67p\h9xosfR`҂Mbz/g$TgwV  +ٗ NmnKa6 n:͵{Ԯ\}Z$$:x.1"_L&z&{z\y|L<'KF7Fğvlf\ptFk;uCGnzwz*}#co&# QqҪˊ;:_ۺ:됚Q5Ijn*[UUV"dZ)BJQͪ!Ѵ#V[RGj̥ڮJn|X! ֯Տc7N)/^\uA(B7[*зÒK^m IhN)U䕺C*2L62@}j1Ivf)"H*JPEFlJQdf_JE%tLzLuLay?}fAޟ~m] _KjAo<cc=&_X(6&0jaΔFa6u3R^?1(4O@q!fhMڙǬJ4e P2OuOuREZ4&y4hJ|!hT?C7g2fd`%ǯXxIP@2~ZhS?|JLMo@.}LPIE+#,,, pv?%ґO倻++VXxP xP#1!ÀG6l= } OS:Wek^`+``;e@`' o6`]{Fty|㾀Cߛ}<| -;?~ݵ@GU۷;};{o?tޡ Iy \˴"sFzD0h_#:QVt1㈞QuuG:Mܣ$|zUU7D5D!H=idsؓ$~ۯ{+oYfٞ7= ;[Ȱ#͘nyoeGCy ]`y͎fԀ[ <"(;Qgu:t<EKo[:y6;͘Uvt`5&S~D`J6#J~j< ߛv#*u:md}©F4ZmK)^+rܻwjsnKT?x- rsXyWS}JOdSVfJcl16=0ɞ,K4ޟ\$::өaԙNLu6۹&rD7lt'>5;˰R \wBizlp2٢3.+\2]Ԋ c''= n^)O ~LY})yb9v~wE7ΚG`dlJoJq?֔IRLiw=Li3- =Uq7? |/3keNGYl'$ԛ+ f夨(%&,,dO:?Mr~hBOJ=([, [ؗ^/&,&NV>Ӊ3h~-##Lgk`بvXL<[x1J=Yx^Yr\9Xhb7])K <Ո-+"4{OO4[HƟ,s2z5 L-Bra<3; Sg a*zw!3mɸ$fNTVZ3꧊ƕxL;߮gb$ZI^MK7-KeBZ` W[$(OV\r30d-Lн&|dF؅clmL.&)U@$}wg;V*XEkzW:?CPJ<ճD,6B6~'Z8[m50R2"ߎVe?5,x:% y{) `nv+M]#°v [Sjalid؜M93>>RaIҰw3~҈ܙo|ij3ddrE6CAxIJ,c&M QUc HB r v3]("srz7\.ŔZMdwb_dvG^H>S򛈻 y;-^9u2X,kN=3/OD|Hoֳ@gPDBr@=| Jx[ymMm90ǍG\%Q7I'5KI}/bOVl#2z'DOZC`TDJ|ȪP~Q:/+UZm*R!V *h W-o\[.jG#WE-fc Gri|md#,k~T2HKȾ u"4"X|Xb3~!r: /c[OI~s *j 0{½Z\[Ǐ:X 9`1 T2%>,s"APT(lO22^e~̷հQ @I"R~Z&"Kcb-k^%}Phڨ{?t!)j@pŐ/mdo[uo[uooa{CmHGKh珂{Jzfg!z}z%̐,zG^W}U)*DGEZT:`!ygm2tT&)f}îITk4zbG%?)Q4QgSSa1Mgby7뻁YkxrsD޸ܭy<;O=[g yq^z>xkܿ)[fu0 WT "+|Ejm E"#‘dX/J3uPfv(GC#p>;-c2;,Ҵ,e"RC@'1QX/FEd`”:ԩd[ qs&r5,"_d Ed;b!7Mhw.*eV:3~5Fw=ڥo4_~.u?I=v5[7$Gfʽ2knok\'d2"!"y1/$?"D @9DDD$"DBd(J"UD&RCp"# #y&2X"?"2)D~LT"L 2$"9D&BL"SF,"Dd:&/Df'@l"9L<B"?%2ED.&r 9oDOdDb"YHo2v5}WDMMD~Fd Dn ,OGgo4?G-Y&J{׾ъw9FƘQ',.ߔlņ[K,T?_z-$}%Ր.nn᧸U#2ZM,2?m=_-$=YzfR&lw]Ag=|&fnwC9X |g?3~9/vjoc;%KiFOt{hخO^.6$H~wSvOqzƙg_\'.C6 LX>t5K^iw"KOP*é&K <3p[:}#B={E VIτ0nrв)IO?GF%?g'_ouQ@s7~[X Mgu^z5I0Fp݅.\nw {tbaD&"""-aDl5B;b,Lq]1j-ʳl2D ڿ"2 XgwԠܑZ{+GdDL!2eBJn.\nwp1XnwMRC֔\snub1Ũ3gS[:7uZK ]:5PgIt8<3I=J<=c2 'oP_,Yˑ!zm(YP J8A{?Yo"k_*GVA bmfǟ`O@k?ӹ?_owECQ'EV뉈"yc BPViQ1ujQ"Dƒиp7gwL^; ZWР|[CDdMJ׈H*V&0QQ9bm7/"gV]:1/b:1/Չjb:oTΣը3k-'0@[D}?HI}5|1G^hC{Q٢"Ej!W)ejV;k0 ISW)JaeovYUc:[*YQZeF5QǺs![Jq$H) h_bF3H7}MY~^<.Zi|1\!|*@OT"j;MҬk4+^E.LVwROq3BW[)ۍkh 4\,2k9H K9;#A7v:ڲؑ,H[7Ps [q0J`29AqXI?gOp~TɗY]y쒗 *]rms)]"ZXACA-ɏm*BއUϻpÐ׭׭LneSN<rrwa^"Gv:}R0Ei69'| Dܢ."RDdo^N':Yl 9 %VV^"BXD "l*@$u;3Pb}m>HxҲv )tR\q#CE$s܏:Ost?Kެ>Թ8:Pg~)*CHӭXp݊?uNکZ Cyb9'DXN:΂\i-DPG+Z:=>q7Hf7 sgVd*A֛IU!$[am1AdՄ#Y#9]Y"A}+d@mAd-!d\@!+E2d &^u#Pv[7}]AL~U)"}"+t #+BeQK'|nܙR<,>Fm9󈊬*bdW!V"AZ3YY[}("5;Yx7V l1eEȚUPu_N.HMSIo%#%':I9owREyw+Wr& +?uZ51X?[>wl7n'|0dO,ھF¹D"3ZύޕGUwt'!! 8xFFE0(BQn1( h;32%#H5 +o@G 8Ê(Πٯ^:N#^/_U~G)kzwp|`-{ÀuG C`#qz=O u?ΟEK  ` @*5V67 p;G.n| دgns@;?Apk7o#cZ  [G\7,i @[~!@!3@BuJ]a@Q==>~ASN0p: ##cg*F~ 8p|ݛ 80p` 000 p!"dŀ)j%*8 p9 tp0pj,l\GrE<+Sx*=<퓄*VLӣһm+7YWAq142-q@71&ktp¯@)qYmwVgJdH-̱I]K>~[JP:oQzo?eo1 0G}z,gnC0rR/rmR_vF9m㗫俫֟#.6.]Nۓi/1V=D^~"]=z#%( 7)_ѺŜ{}1mT)e@cCh*_^RŢJ16|`;r0cq`j#KTقFғ}]1QD_W; b$t/?Zώx cyADVCr"^r)|Xϙi?ACc'uoʳxɕ7zhVm3W%~ߖyvܾ]L?s˴ ݓ%ېe&_;ݾ!JwH!So@:mRS0g$_L< 'Y;4z6/\fg1tENp)+)NCso(e`=;Ȏ Yyv)ǿy^ϓ(~ͶS)}|K{Կn$ԏiBo:~"[;k~kz;uٜg2E彴o3V7:-*NYH{ gw4Xp y^t<bz%=:Yc4u'J=˃yUT1H<5(sC'ŐUŐxŐ6x{{_EhVv2/dDnr'NVѣS& C^\O;3,Fv;1"Q)8Z* 5,14yb3Oȋ[ynXF*MJ~4chxt(#E7@=]Vޓ/SSjJmȅD}R@n]"-A~?:)G8eqܧvǏV5~hv\wMlí}t^NI4ϼno!l"S_gT^y^<ɭXJ fT<92fqJd&hI#fJIeOfQIX ?kKKuw-Ut# OSZWb.qw5K^D1I&b%Jc8_RC*O YNݜGq$㽢Mc162'{0= KrJSRa| oQO|~.;J6Q `Oz5GC~N˦#lW׹Xhpx/c!+_'I.a N69\ A@(qgg@P(ter@t_=\>q` ``!:0pCLYQs? 8O=?NU Iz"8N\  K5:eB+ ?P !Я8@,\XE7Ipٍ- p3R2rQ}mѣ W޿{a}GKlSLL31l?SSkcS*ZP,ic/(f(Uk-<:W Up̭5/yb`Pfn?o C2[)뒆~bb )~(`25|Ryv<=rN3*zom}@ZSZQU̳ǭ7a{ 0| %gj1j 8gb1gEU%xyo9r=iFϿnm귳nچadg=u|/ oɹs7%_'|eGZboom`f-[EbѿXo-[+jCբ\6Ja#_G/17oNNNNmrMcccc#7ȝ9NZV&$֧MJ-REaK짰{ ,r^R 5}RMja1݋,]骯2L_MeE:1.Pf٪Ll՗X9]75 #fCkwZ]}A,FU>צB-?]{v]cļccļc#LkdbA+\%UZ}2sSS4h4i37j|˵صmqaJvo{­&>VCC.!IJ%$[s4cu%LL7lʪ}{ǗM-2mִ<[NAaqak8BOzjoAA iiY6TN OEұ2R;t"I&VM)'NC'ISFҗ2|~Bs: In-aPFb /Mhn@iR=[rh:ա8#v^G|A-FBxSP"XP*6 vS8$@ѠVGO- F( CZNhjmZ(|biC@'ǽ< !1a7 DC6o%tJCIR>h_wCb٘Q.QI&d@ FPAǂ@ 'ZPp%1~P pnWcyL u/+dtR@HPA $Br`\қ h(܀>z4\(f Qb-Gq'k-7qxU:HEvW/g$y_!^b:)g$K$A [/=mZG6xtQz`*DTp%PzP,>v\WrE,HL|?? 8}.$X; @~(fڗ)A|6I} cXM P6yiD_eGk;6qMT-V<ֹX#uу=eo7Q'?-O.;d0g=joQeo3!EYӣiN^] .ɨ)cѷ<41ߍ|!5qy9I2M"rq>Ϧ"6d(lG(&bGُeQ8"l:Qp2Q΋E?A.5.;iA{3 Y[;.PHH{ JS?chGܢwmI:֝^G9w%w8G wU&T5025wð¯B(eإa!Uz!A,֑`]>dy[  zTnTVdWzA}{Ag!\]7zmX=R1 ؚ ^3zޱxUp5Rc-;wy*8aͱn lz^p;C4jb6i=r7>CnR!|ԝ=ًu bH5S᜽ɝ W1}!Id^u/|jb,\VPxM-WE^μ!{%tOǝ}Xp<8 }{ӁW})W~mj>!K8 w)DZ@MsgӂuRG7%q&N8jtpSV%Bwz3i?HJ2M]O$F1nDG=L: Qlh?-\ w's'{;;XO0yi\w'DZ`ټC)j>!K80ko>%86S$ٝ}٥7|t9YȑIΎtA؀SƏ|Fi3Ms?Ƒ|>^m> ߚylrFB4 Sxl&?OFr/\ 'a%g"&V|v 8Y5N/[Hm?@X3nm?+6 v"^򊝖QizMΦe]?$gAwߟ`< u{žr?rvOQ>:]=(UM^sT`zyaAdu$Ϫ2j毃~m?, N6)8tzuupʲd$E+gLzmEYp)KG1r5SQ4>'O<"3˔ 6`[V&F/90M+ -̎Z t'=!qn X}kz0=l_[EE o9Hy ;B:n!4R_[HDBO㸗\ R#y5JY͠ W$Q/!{S,R]@^ B^n'rw_U^Kd)ϰ䍗_2e1'f%j2\L^۹s1?.Tȋb׹xva~"\kˏRe=//\i9WEPRiZͰeM~o.p H'Ӿ9zJf;@>lO;Qgbs3 < z# GBqV9`|T:(2 BVZYup欃neՁ:e2ɊLRMI:pSVQ4QV2ahcK]=Bh$(IP.#MrҴsxnb"Q͏5]Pg&t6=]׸knԙ:nunFJ!M&mY)5!,rMT=ӤKJˮ7v0XKױ4:敟_  ̐ψOӾ:iP8i"8GlUW:Q ,kTfٻs}{$Ӽ ˒򶰶}ZbnioT+DUMviUlC*v*փ[ŶP'so/Mh>%K@#mbB-bU^qyg_rkiE;:+A?Cs9Z*L߯bu{x/{ӗzIyY:׿e_71}fNw骁zqr+8;S~1weeǜiͲ/֫uyR [+,V~gy~X_Ϥ_OY|a?'Yx(zra;1|Ɨj>9'A(;T''HP>:3 gx;sBvDB. P7apN1{Nv3'~%Z9='FK)/'TTR^kR"kFVƣ "͜ ڗucj.v_ԭ&tei^LW|S oߎR6k+l6Y c/2Cyet5E"ߢFD߉Iθ +aO.tif'if@4B-/i&J4,iJ3M̔f4#\iI4H3(vl/Y;{߹_nNlh[&Fr)w뭏:l|Bfx TCqnA*/6m)L8k/chqo*w>#]9Jnr(rBÚzU ͒­Avk7Ԋ{ۏ4B#?#3[5hniek'.fG771,1C &))\TՃ 1e!R-Th}BPKA󛾒5o߿5~[{|`/ϑ{,ܤjqIc+~8zNf= զ˃iz-$H)A1lk}+#K-Wm.t%- Z|> \C!>!'AU_tӗYC.gs|Bf{^%̹=Au8Y弙r2Uε9*U _lSOQdv G;i:Գrmpt@hv*$t YφdC9. ta'g9}PrL_e%*?]*|^ U6˖l1QfkvMsyϦg73xuOϑIHH@@QE((NJQN# ,rx"wDo'x((DT@`M$_gQuuϫkUD˜*"Af,0-U/ˠV\Rj%-VE 8TْJ6 ;eMm[v!It6S֧K3%U"F^.g%Bpo 9,2%TvLj"\a9.}9_7 iMigDX)%\d?UNOI9SY=m{$$(/Ř!rT1>E.6L*6$?nkKwyekP8~um`sswMMzM]l0b5%-ĵifj~bV+6ɖNyLD#LZFϞcjZ8 \J׶w LsGgbu'Y5(V'h)O.oB%?gHI2b,}Bz;YY(6Gk}9LwO؄i9? &kgvVKen5‚ܓI%s\*YޒJ{نZ^W,h}"GYT5j]Ԓ%-#ڹTri!ҎZ؁ZLMֹ&H\MXM=x, 6p; ` x,{?2pX+J X G1PVIX Og` D|_f|+|F5G ~A&d^_>6W6obAPo߫kR3YG{cC/u~V[cF?MSo_p*H")u\WBeX}~^Jn P1Ƃ+MP{~$9@ˌ2Mi%8NP8%%՞01]( LjyV9.,%[ P|QѪ.uϩizyDx Ha R@Q&ݦ4>7k69hZ< ֠mŠ=(@GP :Π t]ap #Ap8ǀ^7}qOxNId0S 0~=?~ gQP`  Ɓ6m/6_K_er0LS2MJ01[c?39g};{Cȭq>?ؼҼ̧]tZDV?w#uң1(==q8g4#uJْ#r/|V}̭ 9^p+8L$*e*Hx%8LQL etJ(S5NQ%< Ny8uˣУ+VfA^_s+_iK~~ЧtK=A$s,\i[m)郴)_ +cبI] {GK1ɘޅQ޽l5lj/ݰ's\ {3w5>ϟͫH/zD__caf؏׿L+ ,^ؑW#G2bn#q"KzW"s Lf|v*4Ofm5)x߈=l7^.:{ڋ߲=LOCr>y)`{EM=+y?dNVgT{{X>$㏗e$VYg>?pg9CMܟ,#b{׏c?x}IvoW^t&KFΜ%3P{镃\/Ɇ֦>e˔@'W{LkaOe93S<^b8[1aۚ&*ut/?WLSgo|b>KK[vWYdyȚwlV>`!#ܺE̓W<^*>Pqe~ͼRW{ ]z?ֲ1;e0CFmwBBl?d~8_jӞc;3[ZϣsƬ;( ]-F_%pLe1^#{@#{mdǦi{ɹLy^VmG ag#)ex]K[G9?v)Ѣ/'֥'5c;_~Ǹ;?ԫ8umZ @>2ۯ?#2W5Zp FpVpwE~~?,c~lWJx<Gc ' ݧl6m5ޯ`-xցz+6 &o.xox1^?&9/+5 oV-l||vv{`Zlg`.iZĶoҷ>U<9] }{Go/Π̏#ԣ'߅"6Ou1=w c].uiQf/t*YLJR #rtk/|}KJNiB%K$qw,"?v69Rgq,#'jyS.ET;tJ}Ԇ~ȓ3!ZI*-Aˌɤe&2Ȇ;o l׾nA}/浯iHSk]:ժnBnkJrVjږ#;FYQJBIT%V |G̗TmK},,Skztu߈9t* gK't6޶3P*_l|.&9( ƾflˬvhi" OFo })l]NGX?QOm%^nvA_?M{C1m+[tUO:t"$CL_/jUvn?sȩo#s=Qi`l] k%=5mA"bPbۨWuюeguˈX_G h㗂![>? $ 4};35d-V ߐt m%x:R teQ0iȎѠ'82}\t8!Oc`Q~Mo0ޟ 0gu 9 0p>\&@%LK_er0LS`f QgoR֌/YZ^\pQ]u^>}]_>=C=s>͑s}k 7WgIF,.Kr:uYc5[5)*S PIB)I5$/-jbO ɤN{jX-b_]G5&յt;%Z$*!)@%zD~;*^J\ FnRoGurJ6̡5RⲮd3u8Vfjw)o.lmA%gQɾb)*&~ fR -b)$yʭIC!ݦ$٘+YbZٮeR!qihɝ˜ZQG: e TVRS¿-j}GfSPKֻ֩]Zjx7?] xTU~^=TUJ%! ;.HF[EEA, MET*aQAF#-3nQQ!tܢѱiL4jRΐ'?{rsXwa^, `yXfH(ZxοZ x1sI3X/1cf 339RꩽWuZn:Kj. q`fqHTeTUM&5- Gt[#ע?5jQff&hc} ȏMyT]xW~ rggrGˏr` N'wtZ쳧;XGXCg%}IP+3łde>(\:wAf~F[~ M.ζ_1ϼatĜDt5ieo9IkմW۸VEVl{]wb^0Γ3[7 ӝsCs,ܷ'گm煎xX\?P&z1&&щi{;g‡ׁ5=z!8oo>mu׿OF?3E]:%:bFcXoq]$ocֿhᠱ u0(f[̶aW3υD'gp_E>*Lj s`̪99?ƽi0Bb`u6ӟQz)Z%1KE #jyz":6aWkZZ\տk,Zf)R`]g6sN MEZDzL$, b;JٽD(C}D`>@_ly8bR^τήg<D$BA$B'֍e>cK̀Y`-.2gмZZkKJh$6gm$$3{ 13Bl7wL [P4tF]$xBd x{K轘#T0̈",S^e.Cqƿ#dE(8_1&1F1yXB2ZIo<C~#\|DUs.݀Հ#ܧ[8o6l<"> l< Oϧ̀#Y6s/^xh ``/`~ED!poϷmw |_|MHw?QpGcsq/_|{IS?qv'zlZƿ=۪=>q7[>j?fMKӿimXu+Z?:ےdL+{Ѿ:HcaN2S-zۄ4z#q&C~\ϖƏ9TPwv}Ho;ұ~?H:b'M77^F]R>`#_e:.bJ5@W;5*E{(G&|52~K-{&2aކȄ#ܗ9S 1$>6fgu>zz$Nti{&˧ɱ(NrrE/gT\?NV~XOgg³lMD<RrMH^|'t꠭d ,r5_ij$ 3sNM " UtSNWp7o4g餡4'\{` #>8]}^KBڽZ?L]{s.XswXؿ>=l7Z+wEKƿFڲ~>ROQCªָzxlgOJUSkt'|s!*1LA(clM[ت\-m.ĖaKKFjK&SiWmbd_vm̃hڈbf03\ҊsÛfT3+ k 3gbhfbjf<̬R5jbtaƌ }rM<31t˲ۙvz{K Z+8+PgfqtFsG.f9lR烹 !S\Yv!b8JS6k3 C̱iq,11bվ8"N5K51QA95IFߩ;nö23|Xf7{߾Br8-OE^! ָ WZamFV,1FQK5+ClN]m稥gb/N}ueല 8KVM5L50|,ӎrgY!c_r%%C5f2׎e*ɓL.5/k}wOcg!L~vNvK k{V02y!lb k\iq,G[M-٬j6<{݂-X&Y%j:$X"ɤǻ\ǻ\O Ye8ַ8c['c}5Ùjj+)yJbkٴ܎1Xfbg`;?^ Lr|h߆d\-e'kqxeZQ$S9q?5rkq@Qdjop ;g,k/Z j3e39ks.H{KN{/S;]qa^YMK~JEdp]@4J1|HPX%/Y"9 Așe3OL>9$4-)!qOraQNb=3ѹ~=w=_dF#9s~?U~T뇬ĖEk>w|R\?8]:~@3eLMKU?WS!{ Ys@ 2>,SP3ZP'X 3~YeH9LM^BAב7حf6ً̘d:o1&$7Z>+ݕGQLϙAc p AqT@kX !rVDX4' ,NQ}R;^ׯ 5c}cbk4/}~9$ű&Xl>ڛSkoNى}#/53zޜTO=>[Th} y ěH]41GINtA\GZ*̍9iNC]0S꼙սPy?]OmRXH텂p3tAr|V=9-dl>'n?TtYFnGgHco /rpm_ЂcbOoRBGK` Т [n7n X$sWp7_<x0 8 v<x `Ys6mu/j{{x&)~ x>AǀOi/-_|{QxhcM5O Bf +U:hZ5f66 e$uZ2OxFGjMa~$#oړ=R/g4y&aS?TP] ty,nt $iRqۙsҗ֥/m >T Sv3U8 m{͈ht~QT}RKgX,ڝf13%V`жQۮ}zaz#Α|LqIߐ߇j7zݡ6a5w4IJh 8O7:hkXէ3O5UVGK7ʘlo)j(BwWd>s3ΒJRKGiH22v3]ٞb3|yn!%4/wXiHg0n}5=;'|6gkS9Si^Le *z$/ˍ֣o5/XH`&вʲTㆉ}Z!+_H^H ɔJ;%s AzȜ~8 Ҡ沚 ~ӓޓ$R#;R}4u='4I#Ɉxy [hNHTgȩ1hs!O#:?[ү}!it&yVnqL$i½:3&Sh]d 2MRi*2 G'Fy6lv8(q1\+f=9 Fse4hkwѾFT2^yN]ПRZE̚^M+_uO=~"CŶ>7N&0$x*M2#5 >Vb4a=X6=ixxzglt=y~#Zyy6c^Pq@NQo ZO'c2WvM#I-6cVnjG m}i1_gw}q3OgQ77j6Iڙ2oPyasP(|չE  p;PTzN  TN  8Syklz;APOQ-uGg㿝ǛwR"=(m r/4r婌+OVH:`+ \q-6s!U&⋇%^GӱOFeQnHC*W͸i}N-ɋ+=[X^DC*U)*M6m=N,.(Vw3+>NKcJrq=y#yRqVb[R+҅ELVZiHf3nFqrTh=ݧD˩ڿ\4O{ѽTvP>^~vt~XH)c^I{UP)[;NC*øZ[sc֮ N=b൴5!iSI"\;>~KvVNJ)gt2' ܂<-K Q'r"sޗ8kS/بk!WM5*u ƚҨ\S$Кy]ō}dM̰;1s}-NeyX& Bv52W2֩hݽCeZZ;X41P2ǐ.b^E z2b_^X Kb:3ƪqkh_/C~ujS-^+_, /Ú5l~rq~8~s[A|uqԔi;櫞H0}cz׷Ow)Kzz Y7v謊Pfw%Of$4>˸\Zħa{+x&ZNȰ2[,H[)cvȗ(cmh1ŒIv)dC_Lz(If8pYt(·Q::!T]/e`)T#Z= ]13k]5?U5e+d>xX?2ҙց*aCMz(sԤ3%|m碌Q[6)cԍ,'f(g k<e~1&=TE#?7)c8f2 jWF{Áq.I'+I/kY#u3Z'}8RN&dX,布gZ8VHsǛH.;*s=q鷻yL'o{Zۙ5Xy>#%g}tPloXT|ͷH_B{2שYFe+ka;̋eH2 LF .e[Q~X y$DG*E)cHBJNH"់6;5 ҮV;*s_22fh [>qsz˔̄,,s_6 ̂b^&he7ٜe/Xw]eX➼Lv׹.,sg)Ag&z~Cd!ςq@ԊNG,iTq{ߚ@h PmHwiwFgZ QO҆'ɛ3 ō).xJ7}{0&>;l]jA}pf.'8FOְmXOmU8sq,ǚ҃iDV;=yGWm^$tn^єkWH:+ KeZgm Wj|e294fq)XEX?Bu㕹+|i$Z6I:_뚙xf3XT gF Dszdlׯgtrb̕7{ys@oT >do>̽졲'{^?WWJ_;Y;N{,-/E\^f;M6Y%\<&+]WȊQosۏZkG:~HgJ'_.XB `삙 2.1,RdSm:+/JT}UG=tf[X/!x=rJ,cnN2رx7a2OmOBgZ\>Jܷ)Εٷ)cҭTd۰' L|=b>̓d<וֹ=]1t>Xq\BV@=K~~{[';ǣ!ωEF^/j`2nڕ禙o?cejOӁ@ҁZfzEf Ez$]K. +7)ԯ}+πU[«{ `-.:_w X& AC-ޙ@GQl}'LO2$ð%$>pCE!j_P)XE`1KR`,߃`%(j_?u`=`kM?-~;.3{^O s!pRp8Nrp2qFٮ0.`Ұ]u jtzY™d\j\ENF"|N b"D:sMgK8ObsF_6ׅ,gTr\]6!U.͈r ;5rBAx=$E}_K%˵vۊP2Z_1[n;5 =~n⿯>,YCCŘ/3| j i_|^R]|%_?Ϗ0+q%s!Nuu[DPB8ϖDkJ%ZҘhMc^Vxy&&)GiV(G5Ynm#߹LnKP%nsκe|k&@<.@u :#NaYIubΘ8)ެn&[ :CTMu2R$sIu5:-ΧTh'Խ&P\`*A3KGQK,ձśuҭoۨ`stP39|MuZ:$ x3lIu:Wկ.Ruۉ?LzA]p@4#oF[ҭL0*px4ԈT{~jx3{p*y*&q`sy&;e|yo%id07ꤶ:7U%uT맭5Y҉1џG\L_ډFo1s S9o_)97wy~[SgTf=He=E1h&yUd+ՙfSuj^UYP%VKg+ޏId1ASPds=lrOq6ٜETkZw۟z--CkGW,'.^kwJvGvGu4V?8bhuG |>ԩeiC\<[QZ/qU{9G~[y2~ ^ڤ{+ R?-׶X ѐȘ<]7lK5]T_G\o… K䄅~QcʏMC$"=ÏAH30mV|髲d+B^{TnM `soFMa.Si7U[0o79sy_9 G?fxH糖;#"kcmMSUiSi6lR7K5'M2~21de>ގe~VS;{ {F;+Ô-aŦBP΋MQF QMmV$?GvamݪDse$i᭏gu|?ϴ knN9մĴ+IVNނ͕R}k^hs|m=G?|#^[^ M6eD6eDqN:pE6uvf_M#M}a rxuR ߩ U^4Amǃ͒xKW)}|~RԘ4J::S-ژ%K`$[fTȊ6dfxhOuJ1h=P43*w%U`? :wST'5|@u~xW3R-? 9-P<YBu⣩KSyͨdD2 zKSF%EPN$Y|HfNQ$\(POC(Yhj89f{B*0K&Y#f;dnYA6Ols,Jm^ȜNqcjOU,G%F{\YbGحllD?A7li>VMi%Kci~w[`lN{H3M"Ax1@h@68@0 @@#7 ăE'ɠv~se;4 m@[׃#:+qt5}V֨?g.VyOK~U՟FjJYT|̆l[*LgޓϱȻ\OFWrM`-/ P^| ^FY8XP8JU|$?ifaTN%xUux4ݒ1{Y)0Fz4̸ RIv[hIsì,5Epf.fy8_ir?߭R*k7JKSÆ\.Dj e7r׉~Cw =6e-.:&tf*#sDhv",g0b,>⯽gҞjgJ$_??p.[;xWp|߅-R{ 9rTfZvUv}N~},ud<֮oUkW7; { #ʬL: kRwBnW R?o.%s'rkyKPOj?p8HST^-uY6fl@!Xv=`/kp{S{`*\5EWP b[0ɌuC < 0 gA8.J.s0Ly` `Uv*Vvl϶eی͒hrEVf)"3 dW,OG0Q*˲ZhkI IoJzYM/\.`.;=QT]&>uk#"qhhX@ v ٷu$h %1q $D ~ h9hA Mih v=tA't] Fp9@e}+=4ーn>p?tx<2 >?O'A&Og@Y0 ρ!`hlpG `$x [H/c=ƀWA_Ϸ*^KA>(ZX(G1BKd 6M8 J@)JEy A Y?^l` X< Nrp4gN8;8Md2ϖ?̟l'ǙtuquA8꒮U^Vk*_U'_RL3CLԘ==&ZLv=&)c1*cfr&1= 4@s"M| :. ĀF ă R@*3L0 s`X P @5j@֌ Og@ ^=0A>}Zvۛ(+ٷ WY+ES$)s9ʵP$Zn\,0 qO ?LQS1c׎QCҨV4hD㒢i\5轂uf:v/!mP׃0!D63L0 |怅`%Tjp Ԁ^m)4xdhs`(ksb%;wҢryk9XnJu)~3] SW~;g\*?-ΓruA=j_+MqW,cCILy0,zl \~A:xL9` X `=68 JP΂ p$\:Π 0<a`8~`'O{*O?V XAxˬEJ21}-3c{”JhZN)jh9$SO?id|KCZ2:4&9K`>70 ( V` ೎h`1KR-(kZ H[@p}(+NS+]m"IߊK} H))動?|1*4?dI/)=~A!uۢNg1^# Ap29`:3,0 "JP .{Ao` Q%0B,y88ʁ+_@`6U*X VRToDkDK/r$B;%/bɽk}JjMߐc8Pp@iYrȐE%!X-QYg1"Ah@":Π x f/%6aPJpyP uc`X<,lG1p')p@hXYj{E. )H"% ^*Ji&!rmr]JGE@(AP UC!EJٻ*f>3Lٹ ӅS`1"ijeMX:~K[r}\2}=ľ2uߡ;r>-wЃ%Ss@ f0h0)XVv`%/@"Hƀ` 怹`X` X ΀zH58[(;Ӷ-Y0|* 9&zl.pc!(vsIԼPM1+ֆ M)מGemZ.g&P>cX0|6-`+8c h p"A@ HYpDPc80_'Ap'p+'/n[vv찄Ҽ,̴MYΗd7.m^r JjəGv+ "$BrQ<sbLy`ߖ@ (Z6 ,`$F1``/ ". EA bAG .`X {Pc88 N+*(ׁÆI*@{!}.b_Lف{ FkDr5ꞯxm62R4$Ezw\kf}ˮ{ÞR1YGcd  @G #0~ 7`Xe`9A <'3x >I0Ls\sЀAx.Dt!<}odkf&MJJE=0z1"o2ŕ|m)}|w͗ٔ#uʸ#K}zyl,P&#,KR \:pEOAuPA:XAa`7΂"p% nuĘ;!OP>h@A`L|eW/e{ܙs $>S*v/KZ['ɞBWmulԑw#Z/K+(Ee4&!$ [&`6y`b89&AᅡoP0&i߀НX ր =6(8N7DD`.BXF9 Cg`:8 N,(p@z^7ЂL5 gE<hF9x֯߀h ,=07(~ZOⅸˀg5 A֗T:?H5 W\uTtb~\@Q:̍3,d27z.M,W24*ɈeiR;GGsJ6[ٹd:vHЧUUktޝP-?h"vjɎuu]P+.ꧥsB.MUAFX4֋Ԑ| xYY??hυL|bΘ wwU~h.$I&LvռjxT9f}n$'2d7eg[HQMrf$iT+SiUUgrU ;]]is|ܛl*wu3vR9QiMwq"'5|fLw-0\WUywoۓOOSa+\+_=\OWW{v1tV`F=|{ R7~gynA"ӫtqUƨT4uRKi硴9Zu/N2JN85S V'9Qfu;-b5ʎȳc~ˆd ]"&=>wtfScAޏmOˬm "5X_vktNo a 75xY^[w}.IȒťsMYrJ2Ad+n!x(m*Q+m>JI@jVmC.+m]wd?Rڮhɮ#^#غ)~`X)*Xҵm] 6CհW!7/w(d9ϵܐsc1{I6.fS\ yy֨9}JKFϓHC.JZW |F`g$O{]N%]JrB;b @j 59=plݶt|7[^DZ~!(׶*n[`[lmA1=8ik*e ԥu|N8>6ںo5Ҡ=yKVjuZN,v"3=k;THhd͆l AcCw=uL}j)FcPMl(b1\XFNomNl2c_3r7s B!)1l&6 %v!EʾU R(C״ 1..vxEk`]k*+@I,\u:hrQS<1G䚲+0YEFթ^UCDTCU#Ң$ZpKF{8f-Ջ FʑifY5QBdRBrxd!ImJq#TTS7`aޗjQ߳o$U#i9OJD$U T4&*&M.) yEbdK!H)VFoCjIr}FJok-&:$SjI5cMf ѪR rEMl բ{9959S Ks&\EU|ͪ g}-,xvfXڄ{2`oU}/ {)}n12H4dE3FyR-_lQ"}_ϓ^֫QB^YՈFcA^h$I#\C_I偵,ɮ  rʕvUrr*ܒS.qhY<ݘS{-bKEN(쩷C4])Vf1dZ"4]J4 B_jxB2ӳHCJa&7j6z GIMKzw( 6-!^H@#%=Z9&5( ͔oO i5EZMi:SݒveEEt5X_S#oщ\݄ܠ}fh3 ~-J=ٳpJὨJYB Nҽt"W]{7龪t.{Hҽ1ZĦXoc I~vG:Bc&ueBz;Ak<{,ݻHU]{7%GvFS$_T#%5B2WW1;\o cй`bP;i`V+M7R?2&C\=G \o mk%2M>MZHf\bVt]t` -LX+j0@09{|slvp 4{ [΃pڭ@70 lf[mx->ݖ(OAڿ C0gP0^Xmw@a;NONcQ-b=D~K_`>`Ѽ p :"+oʴ4{sQ Tp:ibhaW!O+`'.+`6V sz͘gv sE۟O}#noVƤ1m|+Nigoq{];- ڡxWbٖg6{_^Y1M6mʶ[->.fߌb;<;)N˲;9v kvJ4+uW;f1=0EIKYHٔCz2`6relñg#R_<2Yh `lVM2>215@.{/3 @: LXN#&!@]`a 0|0O*lIpz@+`>z$ &,)%¾4Aoo"@4xtEM;EL,@ HB9`Y܀/@+0|f;ӠUGE<n f(!9h "@?0cL`'hLp`&4ep\%?0S>p\W]M :l Bp,"Ϯ1 ti A?0| edU<@5 x|&EsF`"2.z] ;y:Kty;k4E9>JNlܢAm}K ߕWhRn~XeÔ:a*ZyP<^e\~eq6׷t{:eZjݓxSvkХmFzRJZJxtmZft$2S\bz^-B2:Fd&%DJKˈOD"%\^ ԙ_J%cxE4wB BbM2,2?ȇ{%F%â?O'9lӪ*MhÝ+uU]NT +XOU3}HVu7 ֓R巶D>l[mqJO^%%:H47>셊*aW><ҫR5x25/=c_$p:L"ZB)Ok$OI1@,H٠ͥa>T`F YiOəۿ4VC-v"Rݶnٯ§`uK͈jDQ`/-(HQ:c6#!}\dtF.I)Ҡ!:>cbDRT7!00%bæW#_1׾HTz8}N**{wTN`-Z;?doNӅrb?ׯ߁, j2 pD3X*#Oͭwvەfw CFnC sLԭT.sc{}ɇl 6#M Ch -ڀ800ǐ/EC.0۵jLJpJD]A"ZrĘ#PwRc: m ۄE'$ 0h[g杙/d4bRңS"_((Y*+!p6- 2DJCpz[bNI,gp> @ܒɏenV)A)MLHyd RD[LY̮+H2 7fMfoTޔ'V;Zjc!]T*<S<3' =\ y>mA4,֜GgQAŷRx^QaAbArsuX yZ+WEE7􇵿pt`=e jsDFHDܰ9PhZl-؂uX 0F8HXX Huk:USzb5pQg?!G-+i{c@`0f# 3Cvjxސ'Ċ? iUqNxz%BM\ fh%ImYz_h恂${JAu HHk̒΀Nfh Vm_*eO\ҐKY:JQ'01sMB`A2,QW`5ƒ,1ZAAP0@h`Y#MDԅFh 0"e( `J`8*ك2Ie>'11/j1Ӏ4hk4Ӑ7ߠao wM#{FIƎO'|H>L}:fL>5ϛK.KӊU+5ukioMih[ii׏;=itۑ_B:y89}ӕ˗Uy:-m.u\BԥKڵ+u֍^}U޽;eddP=gϞԫW/ݻ7iZ̤,t$"a G999`0P>}o߾Kyyyd4d2Q~~>׏f3Y,ZTPP@'Xzh4h M: ?&L@'N>>c4i}'4yd2e M:>S6m}g4}t1c͜9f͚E9}4{l3g͝;͛Gϧ … iѢExbZd -]-[F˗/+Vʕ+iժU_fꫯhڵn:oӆ hƍi&oᆪ͛7Ӗ-[h֭Ӷmh?Ў;hΝ?Ү]hݴg駟h޽o>ڿ?8pL tߨ=JǎǏӉ'ɓt):}49sΞ=KEEEt9:<]p.^H.]˗/ӕ+Wիt5*..ׯӍ7͛t-_۷oӟIwܡwҽ{;!jg7(D"+r*+җXiBׇmvҿrbEz SJ2M=ؙCi7HQ`pꙵu?, _PhH0hN[mC/{ۿLQ!&:RB|'JNJԔW(suڅ'.ߤ\\'.{벪U>-{P]pHpnZs{68T v[ nbBjEi`9?߼ PK Z!+, figure3.gifGIF89a3  !!!"""###$$$%%%&&&'''((()))***+++,,,---...///000111222333444555666777888999:::;;;<<<===>>>???@@@AAABBBCCCDDDEEEFFFGGGHHHIIIJJJKKKLLLMMMNNNOOOPPPQQQRRRSSSTTTUUUVVVWWWXXXYYYZZZ[[[\\\]]]^^^___```aaabbbcccdddeeefffggghhhiiijjjkkklllmmmnnnooopppqqqrrrssstttuuuvvvwwwxxxyyyzzz{{{|||}}}~~~!,3 H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˗0cʜI͛8sɳϟ@ JѣH*]ʴӧPJJիXjʵׯ1XO`VmA +7"ܲmb݃yXSgvqCv BX%)'.ZW3cK&1KgTdKw6oږ/BԳ+<ǃC~y[|e׳kFͽ|wOc/_}7*_v7b|!X}yNH}Ri˥gaYth:|xť(c+zȢU؈%b+&.#p 9cHcK>9<ףĽW䇊 R)`sTNLFicV6v7v &e{ֹ`'`IhU圌ڴhF*餔Vj饘r馜v駠*ꨤjꩨ*.꫰*무j뭸뮼믮nE^mPfR:l7:gZmPdJk,Bl*m^;%`z.Rz^VoSL EotiX 0MjTZ01t@DX5ZrDS` ;drdYS8r @BR9?΅83X4gjq;_W2oej3aae(Չ4T4U2#{=cVKciPiэkcCuwAjxL/cgQղoX,i#7dNw+jXz DaJ8݋8/Q=Q?(ei?-3;%'1doۧ U;Ҫ ,fU:$Ṳ:&xu%(9':̍P8|%JH9.u#WFXF:NZ3*vCn57 qR]7''bϝX)ic':VZVu kV $eyyTuKbn@2"6RQ U5 IogKnx%︻Qt\g{݈t**?kEsh -.^c 5s#ZV|cBiU`WM=Mؤ Z*U2nv,~˴q]%œX<='j"&{RuX9bmxQ4ƴ F+WFذJfK/a%C36sj- q= Ywd_*0'\2B虴hKR4 ܝ'h2QrT=7׬}rWFلM&gn3U=iY,1ͩ`:>]+Z1kա0vVC[l׫mf9'=4V+8ouT~IBj4Q9;=7jntWel4&޳3.:W\>˾M$6uuA={!,i~bMXPOWwseh{svLP`[B_]tєtC[\ѯ,Vk'/tG5ۏև>l@ P~jmC1],6C^b/\avccM(QYf`hܩɜO<X_:{z(;x99?}G{W7[2$fun~|C{.< z傳p |Tc`HB()2tiV|\!1w0$y@`EI'xe&P e-\U7g|t;Q&wGX=6"wr2"6V7edrj1o%dC`=4+e)N\YNF<%TO]xvߧ䊏6ृJ=?҆vsk"c&8ȸTKEzfG,(sAwe۷t-Ҍ-Zs2uvql6% z@"`#HcXpBdh"BȂĸwtM.8v:f %9n{udɋi%lv採{xlu~X(k·{KiÎ DŽgɆ.GagDv5w_)3ib4yhXFvEzt3h'WFMEs9 BmXu~Q)9 |Hgsn7"sGT]8)(H2+g.C"P9uhnĄGydm$ǚ췏'|Hy3H]YeyD+TOdNb2t;X i4gzyQhuцN㘘%^; KAB<?ɝَEv~;~HMHzQ5\jwٓ5WΦ c:W]ʹjWOreɉh,ɑ֜^>)9Xoՙ9C6tQPbbՙ=cUȞטՔӡ;dw:[ɤFX]|G9NXCgj6$: ȠtYBlw'pUiw꩐śd蠣HB5)j);W:Ly[F|訄kiq~Fx`lz5Ou7BXlE&zHj*i?zC+ifyj iC[XڇFXٚym:4 HEy l@d68)%gSq+*luCѣNyz誊Jh'zȮi_abw{2/C)9i0bbHՏҲ.02;4[( 8:<۳>@B;D[F{HJL۴NPR;T[V{XŇ J~֔`*#%9qhpdIJ]ȶi[ ceKqjf{u֭FJ9||Z۹;[{ۺJDx lʸ xBuWU {F~t$Ly#&ˈI\iR_ʱ xʕ$˽m6*+%\| <\|;PK p[! figure4.gifGIF89a0  !!!"""###$$$%%%&&&'''((()))***+++,,,---...///000111222333444555666777888999:::;;;<<<===>>>???@@@AAABBBCCCDDDEEEFFFGGGHHHIIIJJJKKKLLLMMMNNNOOOPPPQQQRRRSSSTTTUUUVVVWWWXXXYYYZZZ[[[\\\]]]^^^___```aaabbbcccdddeeefffggghhhiiijjjkkklllmmmnnnooopppqqqrrrssstttuuuvvvwwwxxxyyyzzz{{{|||}}}~~~!,0 H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˗0cʜI͛8sɳϟ@ JѣH*]ʴӧPJJիXjʵׯ` `سhXkvڅmaܺ툗o߂l 7׶uUX⼂>x1Ɇ&~x;Ü1x?M?o[7ZwsE}VgqcS=P빼e\|A7me]~(4ƀ2wNl7Rx v V3MtR 7. @~&!L CÝL_R5FH*&k7MUO), ;'l[L5]TtRr^2%/Y fEe*T 32NH+dAHFrz/e(HG{b ?!˒iQ ;pV7BEpL"m)n Sttw!<MT$D(5A=S^ .z C.UUd ʛaWh"][fU* ery|*3Ntʞ-!7qNxټAs]IB%@^mD:K^ҜSBQC.ܨTD*4i#-ӶBQo$%YL&ݳbDeMpijpiP潙Nxx Ԣ|>x4Aʲ.+O)HKo9S(Vn~3~Wr+bZ 1cc@bw=a['Gj6N%DH3Vّo4l'^VO0o(f[cӣILyHiNCo/dUiD%Sg6xknp۽Sin>ǩ[9E_'ӝx׻ίF͵JtC|R Yqu+3V*aS\ *X 7I6'6^ߖjh5;-[Ts[g OuWV6 La4z&ymmLmhڳZ3tI'_ EVRgO]z75go5,a^Զ4LDA)YhaZ huu_!3WlSPm3[٣$ڙ4T-_qKܢm6编+t?n>*jЊ{wָt2'CuHZw=>5'(_ݳԴE;*KR9xYZ k͎blߞGG}BvH{wEzmԖ].k7ogn43wO13jva"$8~cM${\e_s)JvߛNYEJK3zRBek3[7t'p[.mZ% ,h\jTD/l0,MrO~w4v opr,Y#<78@,c}5uPC([1~KEk0WVre j}vkqBWW93X}Qd(*XC?7gc7lki]({!/wQxuD0{NuMp yyi.EgY'&xQHdw`0>1s|Dj%H7TwW϶?{Jcr{w.B+3|x$q(Cxt.rVN+K#V5znZWdtZs ?a2g?.'98(s~ |uU*ĘnqD,N%:!R%St6Hi$@)lYp8SNѨnf|D*vבD- yAe؆Ut0u~ƒ/Log(G@6"bT~1nwP{xA>BnC)?ue2S蓓Y- xhҁĖu&oQ"Lٕ v_,[SU!^S&C''$r܋HZ\7W""@PL*'ótz .6@ UC\F|HJLNPR>>???@@@AAABBBCCCDDDEEEFFFGGGHHHIIIJJJKKKLLLMMMNNNOOOPPPQQQRRRSSSTTTUUUVVVWWWXXXYYYZZZ[[[\\\]]]^^^___```aaabbbcccdddeeefffggghhhiiijjjkkklllmmmnnnooopppqqqrrrssstttuuuvvvwwwxxxyyyzzz{{{|||}}}~~~!,2 H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˗0cʜI͛8sɳϟ@ JѣH*]ʴӧPJJիXjʵׯK@h+4ځlU5Kwho%ɼ~ 7pgS̘Zkޝ0ބ.71^j'[s̟n̚"bϩc_PsΞJ[fM.pعK^ȉ?}sqQӾQJ術zꐣ"IWbڪ~H^)ҬQVL2PrIh}JARi2*z*Y"YzגElޔ_+V%Նkӹ"߅E.难Λ䱲}8o$俀媙 n:Dmko`z1+jn ChaʵKS^_2GN9B?,+CZƑT3 _MFcԧe \DZ;vMOW(sǶ{#$7촋~ rޮwT eƸ2LR0[vߋ+9c8řw'lS7~ޙƏ_$9LT[ϗ󒮇NHKm8iؗk{(ޒZ#apʑr+p<1 [1ث_:EךJINX2(|Uͅ9X'>xw8$xMI(LY"B?%k2w#P~8Vd- Gxkxє jLRL^ Seŭ[?"ĶBmӚ6!&oT4b]G !ݨ41 \2 C tiH+0!$>: 1C-FVӄGvd^sW%aL-@-bN.h_,`Hj$T&'W,-*rcij4}ݧdcӶͽXE N~v7G]}EYe1 ƆS.~|JlS&|(Ύ9J4` -Zϝՙ͍M1!G;ZPe운|[x8M b*=`^2rxܨGCvq([*OyǪ̂j2QB֑l7ߢՒRz" hvT+h|hKhNy;\]w j*=d/ [)#IIYGvLONU%YF98g$1[ڒ*k3~+ 5ž bpeIkܞ[I0m܋yʆ,ɱMЗkۅCxZMz&hil&gRUj+ƎV H0mFe:"V4d&$ egۤW;vlB] z,qfl:j# W:u;aVqABW˨LJN5T{)F+-h6 퍯ʎ/!2sXWrs;R$ؤS5u]} ]ҌR}~Ŝf;FoglY:3?oT|!yo~264sO7<s=y{^wƤ>{3wѣ3.u]>󢳯/y <W:r!^jrg_'pwNYYv{А:rR ~e"HaX$T]EGrK{]yu&6H#Dv794g7|}GOXfEV~. (@*ڧeÀ!u|5Xzbsk9WO (Hjw}HbɧVi5Ez)Iou[\(Mgu_.gd%ok-+H"OF{ƄXYCz(wh`o4tZ<{xޕwd*M/T(68uxgN؉p؆ׄAhf8hQh$7n'pwa26sWaۧtc{8X[ΘP?xo.VɘxW{ȍhN؋e>ևȇШel(m794VxD菎u\\|Hꘇ}Ԩ,577jTf7g78m緑b_ IG(xu8(*}@#y׏6);;bF-'wx.8!e0B ;aU؇U爕I^6`P-O+Ixlow%cyP6tqзr$9pe㒼y{Y3~ U9y*JCsR혈BnH~:D!\[b [& 6ᕦ4I ȌݷBR)B)ߧO֚Dsqy6=&$VIIxGIx蘑ȜAGEGc)azi"sY Eo|d摁K_iZW<FX jKU׬ I!w&Z/aioϘe8^*ظ1]ViP:`{#De,UK;[{۱ ";H jjG0+*~yKha6x!iH3[`=@Bi;=aP'd,s;W_+La;d[f{hjl۶npr;t;PK^!gm 1 figure1.gif?qt|R>ؐȵI7暱\! !Qas_cd2!1cXU˙[! iy|aV|kw A͋DA)k}/ʻK(Ël+i2zA}̌32׫sԝ)~E!tw$msԆ+mr48D%W6B@5MQ^ݩҟ]_vy(T[)O)#?@s'i[|gʶ"Q i]fvn{)r$yO6s.T~ tRgk_ mk2-~nԶ9p˴(U4[@=kR7E;M\~~Wf4tz ޴8oiϛE^eG.9m$=ƌ]Q{"祠ڣ#͵~1Y;7!K+i w}N3Cgu}0djkB)xT O?za[nqOWILtq׻TƿڡPL 6Ylt6?OV-\vy E =Wb!~tP$y#KiEIW[Uؔ |pzϮPZY܋Mg 9dR9G } bA'Zdo<ų6ПAT*M) _¡ N58zYzH)E[5]_Os~_5m >ǓXx=/ w*B1ꤷ>l բ$p0v$y:5Fg)kc}p6xv gF{_E4(NZ)Ee=#NFq9E_Y"l]3TsVGAM_g<u*Ţd\҇5B|טDͽyI;)eb/֛w[E`K3R@F\iP܉]/pAzgs5!` Ͳx2AxG1uA++.dĕ!qǕ+ /L)Hrrs=1wu0~p!/&g{2 y{6z^N~OQNKI\&`ԑJ VuUr S*d "ʭ3v(U%¾# ޼Bҫcqf΅^;DSr1tbn-Z`?zN U >&5*Uބ|d)7e`QQ(mD-fP]'i8Tkixgjj@=+NV}c& [O\_rJ_봲6 b>&\$"}2ˎ(jklo9J}Hu4 $=O_.)zӂO^P^t剂7{W%MxB|V:p4\x3C <ACّKOX]^FLe=P5F9t)}&=IwDPN&Hڸ$h7QBmM%m/)Yf̾uYSLCQ]_1J{<,ԡs}59#;s MۄUB_hʖ9>``F?&5͡brarS%zi~9޴-C,h|2|;8“UÔamONZ|If ىK߳7%CIdzR=ɧy7"@k.g:O;r{AJpGGFθ`ܱ xTgv2IPQ}y!Yo<ĠKᔴť8> }ÕS~ƅFo8 "I V{]МѕH6n4;kz%vsUT ٱ]~CR/IүAAo,SYJQ A|;U\䢻V@GSfsp >T~*7~8A PK ]!0~ Figure2.gifGIF89a  !!!"""###$$$%%%&&&'''((()))***+++,,,---...///000111222333444555666777888999:::;;;<<<===>>>???@@@AAABBBCCCDDDEEEFFFGGGHHHIIIJJJKKKLLLMMMNNNOOOPPPQQQRRRSSSTTTUUUVVVWWWXXXYYYZZZ[[[\\\]]]^^^___```aaabbbcccdddeeefffggghhhiiijjjkkklllmmmnnnooopppqqqrrrssstttuuuvvvwwwxxxyyyzzz{{{|||}}}~~~!, H*\ȰÇ#JHŋ3jȱǏ CIɓ(S\ɲ˗I͛8s )sΟ@ JcOtiѧPJcӂK>Ԋԯ`Ê뿬K#U+۷;:7Ds˷߲d5:bω{*^X1ͦpč3k)u yӨ7F,U M\-kRYoNpZ]||s%K.8ѡ6.yW2݄W*_r†Ͽf_y߁V$l9؃F(`w Vhe p -xervuHbm"6(a`(c٩YiUqԌ<ڍQx-㑜H"ɒHFVk\P$Q]@gdFgWJYPmSo֩_'vI9!Fi(<ޚhX2z夂=jTz^iQ5S}jjPR$zꫪ&`01vhrƊ/\r w>Y2w"C&+mp'h.i܆!J+Zz'쒫{JimK NkyoC:HS.J mKV_$M7^Xc|m ;J(Hkrhfi8;se#r̶p' "vbaMv;[43S3 2WfojGMXlxuv2K5oe*v={,ߔnb'Gm"-u጗+8ٸۙ7)̶9r w@9o.w6oĪ-{l.[.‹&9=n9'lkX⹇/y?Xf1%خV Ewݓ4ݡa|Mtz`ާ-Y0cb }a'4{R♩Saa 78Ȱb3 |*;bC(9Yt˧ҒFwJn$IuM&s 8-NNCTPmV? 8RӪI%\Sa,N޹>u[kBzϡf.e]`ulRWTv i>TXo!fbǩMn4TYn66H[SVlFիRGJ3o5mSU(l^u ҖvmHmZZ.ux7ȭ-Ne2ϤftnzS(+]{u#UUAS!]ӸWL$ *6/կ\s,ӍQZ_Ҷ徎*0wZpPL^XfJ IÍԙ=lR x})Ml'uħ3 ,樋 cY-k#‡<^'8<^rXUaHx6vfd!+s,QrUcw1ޙ739Lf/eT-痍u͉\gMiZ 2Yr"4lfe>5]mjz[YV*i>6i]hc+{Y\q~yP {ۙ/j@sf]B*̑杖]pk{]l{{(盬w mA&gi>zs(A{&6|(OW0gNۼ澭b^׆6ޜ>%mK縘I-tHϽmmv뼹?tw] $p%uW-1ꈪz_7ޤȞ=h;= MEۿ}_į0vw(y -1m<~M+WC:>W{s{e6Λ7*vܩvwo#V6ⷝuwL|Kr_GW[f}]k^*-Χ eT~›>)zoQ,&;sQ=m9jt",7rS0;)>'5?!{g y xYǂ+D}F6(#8+nu<8[;A8B>8~UqGMH{OyQS[@h L!s@ZlRfyw2W #(i#L]pWvRulyhIddԇ~H1VhFc1(i|ňB$s~ZxAUDB+9<5dQn6ӌBZ-+!9G88%A4Q 9ԍPbL0ELb,i]x{ VXԋH6W;5?јn#Z6)OƗ6؆ !@7}[bwDÑXSPO?ɴ4ו|jxz6{G;,<GjU BT_2Vy%!7LI?~UL\w5-ccDKpWٗ3<Ñ ُS%4,Iia*