Copernicus: Sentinel-2 — The Optical Imaging Mission for Land Services
Sentinel-2 is a multispectral operational imaging mission within the GMES (Global Monitoring for Environment and Security) program, jointly implemented by the EC (European Commission) and ESA (European Space Agency) for global land observation (data on vegetation, soil and water cover for land, inland waterways and coastal areas, and also provide atmospheric absorption and distortion data corrections) at high resolution with high revisit capability to provide enhanced continuity of data so far provided by SPOT-5 and Landsat-7. 1) 2) 3) 4) 5) 6) 7) 8)
Table 1: Copernicus is the new name of the former GMES program 9)
The overall GMES user requirements of the EU member states call for optical observation services in the areas of Global Climate Change (Kyoto Protocol and ensuing regulations), sustainable development, European environmental policies (e.g. spatial planning for Soil Thematic Strategy, Natura 2000 and Ramsar Convention, Water Framework Directive), European civil protection, common agricultural policy, development and humanitarian aid, and EU Common Foreign & Security Policy.
To meet the user needs, the Sentinel-2 satellite data will support the operational generation of the following high level products like:
• Generic land cover, land use and change detection maps (e.g. CORINE land cover maps update, soil sealing maps, forest area maps)
• Maps of geophysical variables (e.g. leaf area index, leaf chlorophyll content, leaf water content).
The mission is dedicated to the full and systematic coverage of land surface (including major islands) globally with the objective to provide cloud-free products typically every 15 to 30 days over Europe and Africa. To achieve this objective and to provide high mission availability, a constellation of two operational satellites is required, allowing to reach a 5-day geometric revisit time. The revisit time with only one operational satellite as it will be the case at the beginning of the deployment of the system is 10 days. - In comparison, Landsat-7 provides a 16-day geometric revisit time, while SPOT provides a 26-day revisit, and neither of them provides systematic coverage of the overall land surface.
The following list summarizes the top-level system design specifications derived from the user requirements:
• Sentinel-2 will provide continuity of data for services initiated within the GSE (GMES Service Element) projects. It will establish a key European source of data for the GMES Land Fast Track Monitoring Services and will also contribute to the GMES Risk Fast Track Services.
• The frequent revisit and high mission availability goals call for 2 satellites in orbit at a time, each with a 290 km wide swath using a single imaging instrument
• Continuous land + islands carpet mapping imaging within the latitude range of -56º to +83º (the selected orbit excludes imagery from Antarctica)
• 10 m, 20 m, and 60 m spatial resolution (in the VNIR to SWIR spectral range) to identify spatial details consistent with 1 ha MMU (Minimum Mapping Unit)
• An accurate geolocation (< 20 m) of the data is required (without GCPs) and shall be produced automatically to meet the timeliness requirements. The geolocation accuracy of Level 1 b imagery data w.r.t. WGS-84 (World Geodetic System - 1984) reference Earth ellipsoid of better than 20 m at 2σ confidence level without need of any ground control points is required.
• Very good radiometric image quality (combination of onboard absolute and on ground vicarious calibration).
• The mission lifetime is specified as 7.25 years and propellant is to be sized for 12 years, including provision for de-orbiting maneuvers at end-of-life.
• 2 weeks of satellite autonomy and maximum decoupling between flight operations and mission exploitation
Table 2: Sentinel-2 fast track service compliance to land user requirements
To provide operational services over a long period (at least 15 years following the launch of the first satellites), it is foreseen to develop a series of four satellites, with nominally two satellites in operation in orbit and a third one stored on ground as back-up.
Figure 1: Artist's rendition of the Sentinel-2 spacecraft (image credit: ESA, EADS Astrium)
In April 2008, ESA awarded the prime contract to EADS-Astrium GmbH of Friedrichshafen, Germany to provide the first Sentinel-2A Earth observation satellite. In the Sentinel-2 mission program, Astrium is responsible for the satellite’s system design and platform, as well as for satellite integration and testing. Astrium Toulouse will supply the MSI (MultiSpectral Instrument), and Astrium Spain is in charge of the satellite’s structure pre-integrated with its thermal equipment and harness. The industrial core team also comprises Jena Optronik (Germany), Boostec (France), Sener and GMV (Spain). 10) 11) 12) 13)
In March 2010, ESA and EADS-Astrium GmbH signed another contract to build the second Sentinel-2 (Sentinel-2B) satellite, marking another significant step in the GMES program. 14)
Sentinel-2 uses the AstroBus-L of EADS Astrium, a standard modular ECSS (European Cooperation for Space Standards) compatible satellite platform compatible with an in-orbit lifetime of up to 10 years, with consumables sizeable according to the mission needs. The platform design is one-failure tolerant and the standard equipment selection is based on minimum Class 2 EEE parts, with compatibility to Class 1 in most cases. The AstroBus-L platform is designed for direct injection into LEO (Low Earth Orbit). Depending on the selection of standard design options, AstroBus-L can operate in a variety of LEOs at different heights and with different inclinations. An essential feature of AstroBus-L is the robust standard FDIR (Failure Detection, Isolation and Recovery) concept, which is hierarchically structured and can easily be adapted to specific mission needs.
The satellite is controlled in 3-axes via high-rate multi-head star trackers, mounted on the camera structure for better pointing accuracy and stability, and gyroscopes and a GNSS receiver assembly. The AOCS (Attitude and Orbit Control Subsystem) comprises the following elements: 15)
• A dual frequency GPS receiver (L1/L2 code) for position and time information
• A STR (Star Tracker) assembly for precise attitude information (use of 3 STRs)
• A RMU (Rate Measurement Unit) for rate damping and yaw acquisition purposes
• A redundant precision IMU (Inertial Measurement Unit) for high-accuracy attitude determination
• Magnetometers (MAG) for Earth magnetic field information
• CESS (Coarse Earth Sun Sensors) for coarse Earth and Sun direction determination
• 4 RW (Reaction Wheels) and 3 MTQ (Magnetic Torquers)
• RCS (Reaction Control System) a monopropellant propulsion system for orbit maintenance with 1 N thrusters
The different tasks of the AOCS are defined the following modes:
• Initial Acquisition and Save Mode (rate damping, Earth acquisition, yaw acquisition, steady-state)
• Normal Mode (nominal and extended observation)
• Orbit Control Mode (in- and out-of-plane ΔV maneuvers).
Figure 2: Overview of the AOCS architecture (image credit: EADS Astrium)
The geolocation accuracy requirements of < 20 m for the imagery translate into the following AOCS performance requirements as stated in Table 3.
Table 3: AOCS performance requirements in normal mode
For Sentinel-2 it was decided to mount both the IMU and the star trackers on the thermally controlled sensor plate on the MSI. So the impact of time-variant IMU/STR misalignment on the attitude performance can be decreased to an absolute minimum. Furthermore, the consideration of the time-correlated star tracker noises by covariance tuning was decided.
Figure 3: Sentinel-2 spacecraft architecture (image credit: Astrium GmbH)
Figure 4: Block diagram of the Sentinel-2 spacecraft (image credit: EADS Astrium)
The EPS (Electric Power Subsystem) consists of:
• Solar Array (one deployable and rotatable single wing with three panels). Total array area of 7.1 m2. Use of 2016 triple junction GaAs solar cells with integrated diode. Total power of 2300 W (BOL) and 1730 W (EOL). The mass is < 40 kg.
• SADM (Solar Array Drive Mechanism) for array articulation. Use of a two phase stepper motor with µ-stepping to minimize parasitic distortions during MSI operation, motor step angle 1.5º. Mass of < 3.2 kg.
• PCDU (Power Control and Distribution Unit). PCDU with one unregulated 28 V ±4 V main power bus. Mass of < 21.6 kg; the in-orbit life is 12.25 years.
• Li-ion batteries with 8 cells in series. Total capacity of 102 Ah @ EOL. Mass = 51 kg.
Figure 5: Block diagram of the electrical power subsystem (image credit: EADS Astrium)
The OBC is based on the ERC32 PM (Processor Module) architecture. The PLDHS (Payload Data Handling System) provides source data compression from 1.3 Gbit/s to 450 Mbit/s [state-of-the-art lossy compression (wavelet transform)].
The spacecraft mass is ~ 1200 kg, including 275 kg for the MSI instrument, 35 kg for the IR payload (optional) and 80 kg propellant (hydrazine). The S/C power is 1250 W max, including 170 W for the MSI and < 100 W for the IR payload. The spacecraft is designed for a design life of 7.25 years with propellant for 12 years of operations, including deorbiting at EOL (End of Life).
Table 4: Overview of some spacecraft parameters
Figure 6: Schematic view of the deployed Sentinel-2 spacecraft (image credit: EADS Astrium)
Figure 7: The Sentinel-2 spacecraft in launch configuration (image credit: ESA)
Launch: A launch of Sentinel-2A is scheduled for 2014 on a Rockot vehicle of Eurockot Launch Services (a joint venture between Astrium , Bremen and the Khrunichev Space Center, Moscow). The launch site is the Plesetsk Cosmodrome in northern Russia. ESA awarded the contract to Eurockot Launch Services on Feb. 9, 2012. 16) 17)
RF communications: The payload data handling is based on a 2.4 Tbit solid state mass memory and the payload data downlink is performed at a data rate of 560 Mbit/s in X-band with 8 PSK modulation and an isoflux antenna, compliant with the spectrum bandwidth allocated by the ITU (international Telecommunication Union).
Command and control of the spacecraft (TT&C) is performed with omnidirectional S-band antenna coverage using a helix and a patch antenna. The TT&C S-band link provides 64 kbit/s in uplink (with authenticated/encrypted commands) and 2 Mbit/s in downlink..
The requirements call for 4 core X-band ground stations for full mission data recovery by the GMES PDS (Payload Data System).
In parallel to the RF communications, an optical LEO-GEO communications link using the LCT (Laser Communication Terminal) of Tesat-Spacecom (Backnang, Germany) will be provided on the Sentinel-2 spacecraft. The LCT is based on a heritage design (TerraSAR-X) with a transmit power of 2.2 W and a telescope of 135 mm aperture to meet the requirement of the larger link distance. The GEO LCT will be accommodated on AlphaSat of ESA/industry (launch 2012) and later on the EDRS (European Data Relay Satellite) system of ESA. The GEO relay consists of an optical 2.8 Gbit/s (1.8 Gbit/s user data) communication link from the LEO to the GEO satellite and of a 600 Mbit/s Ka-band communication link from the GEO satellite to the ground. 18)
To meet the user requirements of fast data delivery, DLR (German Aerospace Center) is funding the OCP (Optical Communication Payload), i.e. the LCT of Tesat, – a new capability to download large volumes of data from the Sentinel-2 and Sentinel-1 Earth observation satellites - via a data relay satellite directly to the ground. A contract to this effect was signed in October 2010 between ESA and DLR. 19)
Since the Ka-band downlink is the bottleneck for the whole GEO relay system, an optical ground station for a 5.625 Gbit/s LEO-to-ground and a 2.8 Gbit/s GEO-to-ground communication link is under development.
Orbit: Sun-synchronous orbit, altitude = 786 km, inclination = 98.5º, (14+3/10 revolutions/day) with 10:30 hours LTDN (Local Time at Descending Node). This local time has been selected as the best compromise between cloud cover minimization and sun illumination.
The orbit is fully consistent with SPOT and very close to the Landsat local time, allowing seamless combination of Sentinel-2 data with historical data from legacy missions to build long-term temporal series. The two Sentinel-2 satellites will be equally spaced (180º phasing) in the same orbital plane for a 5 day revisit cycle at the equator.
The Sentinel-2 satellites will systematically acquire observations over land and coastal areas from -56° to 84° latitude including islands larger 100 km2, EU islands, all other islands less than 20 km from the coastline, the whole Mediterranean Sea, all inland water bodies and closed seas. Over specific calibration sites, for example DOME-C in Antarctica, additional observations will be made. The two satellites will work on opposite sides of the orbit (Figure 8).
Figure 8: Twin observation configuration of the Sentinel-2 spacecraft constellation (image credit: ESA)
Figure 9: Alternate schematic view of the Sentinel-2 spacecraft (image credit: ESA)
Sensor complement: (MSI)
MSI (Multispectral Imager):
The instrument is based on the pushbroom observation concept. The telescope features a TMA (Three Mirror Anastigmat) design with a pupil diameter of 150 mm, providing a very good imaging quality all across its wide FOV (Field of View). The equivalent swath width is 290 km. The telescope structure and the mirrors are made of silicon carbide (SiC) which allows to minimize thermoelastic deformations. The VNIR focal plane is based on monolithic CMOS (Complementary Metal Oxide Semiconductor) detectors while the SWIR focal plane is based on a MCT (Mercury Cadmium Telluride) detector hybridized on a CMOS read-out circuit. A dichroic beamsplitter provides the spectral separation of VNIR and SWIR channels. 20) 21) 22) 23)
EADS Astrium SAS of Toulouse is prime for the instrument. The industrial core team also comprises Jena Optronik (Germany), Boostec (Bazet, France), Sener and GMV (Spain), and AMOS, Belgium. The VNIR detectors are built by Astrium-E2V, while the French company Sofradir received a contract to provide the SWIR detectors for MSI.
Figure 10: Illustration of the MSI instrument (image credit: ESA, Ref. 8)
Calibration: A combination of partial on-board calibration with a sun diffuser and vicarious calibration with ground targets is foreseen to guarantee a high quality radiometric performance. State-of-the-art lossy compression based on wavelet transform is applied to reduce the data volume. The compression ratio will be fine tuned for each spectral band to ensure that there is no significant impact on image quality.
The observation data are digitized on 12 bit. A shutter mechanism is implemented to prevent the instrument from direct viewing of the sun in orbit and from contamination during launch. The average observation time per orbit is 16.3 minutes, while the peak value is 31 minutes (duty cycle of about 16-31%).
Table 5: MSI instrument parameters
Spectral bands: MSI features 13 spectral bands spanning from the VNIR (Visible and Near Infrared) to the SWIR (Short-Wave Infrared), featuring 4 spectral bands at 10 m, 6 bands at 20 m and 3 bands at 60 m spatial sampling distance (SSD), as shown in Figure 12.
Figure 11: The MSI instrument (left) and the associated VNIR focal plane (right), image credit: EADS Astrium, E2V
Figure 12: MSI spectral bands versus spatial sampling distance (image credit: ESA)
Table 6: MSI spectral band specification
The pushbroom MSI instrument features a unique mirror silicon carbide off-axis telescope (TMA) with a 150 mm pupil feeding two focal planes spectrally separated by a dichroic filter. The telescope comprises three aspheric mirrors: M2 mirror is a simple conic surface, whereas the other mirrors need more aspherization terms. The spectral filtering onto the different VNIR and SWIR spectral bands is ensured by slit filters mounted on top of the detectors. These filters provide the required spectral isolation.
CMOS and hybrid HgCdTe (MCT) detectors are selected to cover the VNIR and SWIR bands. The MSI instrument includes a sun CSM (Calibration and Shutter Mechanism). The 1.4 Tbit image video stream, once acquired and digitized is compressed inside the instrument.
The instrument carries one external sensor assembly that provides the attitude and pointing reference (star tracker assembly) to ensure a 20 m pointing accuracy on the ground before image correction.
Figure 13: MSI electrical architecture (image credit: Astrium SAS, Ref. 23)
Figure 14: Internal configuration of MSI (image credit: EADS Astrium)
Figure 15: Mechanical configuration of the telescope (image credit: EADS Astrium)
The mechanical structure of MSI instrument holds the 3 mirrors, the beam splitter device, the 2 focal planes and 3 stellar sensors. It is furthermore mounted on the satellite through 3 bolted bipods. This main structure (Figure 15) has a size of 1.47 m long x 0.93 m wide x 0.62 m high with a mass of only 44 kg.
The optical face of these mirror blanks have been grounded by Boostec before and after CVD coating (i.e. before polishing), with a shape defect of few tens of a µm. M1 and M2 are designed to be bolted directly on the main SiC structure. M3 is mounted on the same structure through glued bipods. 24)
Table 7: MSI mirror characteristics
Mirror manufacturing: The mirror optomechanical design was performed by EADS-Astrium on the basis of the SiC-100 sintered silicon carbide from Boostec who produced the mirror blanks and delivered them to AMOS (Advanced Mechanical and Optical Systems), Liege, Belgium. AMOS is in charge of the deposition of a small layer of CVD-SiC (Chemical Vapor Deposition-Silicon Carbide) on the mirror. The purpose is to generate a non-porous cladding on the mirror surface which allows the polishing process reaching a microroughness state, compatible with the system requirements regarding straylight. 25)
Figure 16: Optical elements and schematic layout of the MSI telescope (image credit: EADS Astrium)
VNIR and SWIR focal plane assemblies: Both focal planes accommodate 12 elementary detectors in two staggered rows to get the required swath. The SWIR focal plane operates at -80ºC whereas the VNIR focal plane operates at 20ºC. Both focal planes are passively cooled. A monolithic SiC structure provides support to the detectors, the filters and their adjustment devices and offers a direct thermal link to the radiator.
Figure 17: Focal plane configuration (image credit: EADS Astrium)
Filters and detectors: Dedicated strip filters,mounted on top of each VNIR or SWIR detector, provide the required spectral templates for each spectral band. The VNIR detector is made of a CMOS die, using the 0.35 µm CMOS technology, integrated in a ceramic package. The detector architecture enables “correlated double.”
The so-called VNIR Filter Assembly contains 10 VNIR bands (from 443 nm to 945 nm) and the so-called SWIR Filter Assembly includes 3 SWIR bands (from 1375 nm to 2190 nm). The sophisticated development of the filter assemblies is caused by the specified spectral performance parameters and the high stray light requirements due to the topology of the spectral bands. 26)
Sampling for the 10 VNIR spectral bands along with TDI (Time Delay IntegrationI) mode for the 10 m bands. Black coating on the die eliminates scattering.
Figure 18: Photo of the VNIR (top) and SWIR spectral filter assemblies (image credit: Jena Optronik)
Figure 19: Photo of a CMOS detector with black coating (image credit: EADS Astrium)
The SWIR detector is made of an HgCdTe photosensitive material hybridized to a silicon readout circuit (ROIC) and integrated into a dedicated hermetic package. The SWIR detector has three spectral bands for which the spectral efficiency has been optimized. The B11 and B12 bands are being operated in (TDI) mode.
Figure 20: Photo of the EM model of the SWIR detector at hybridization stage (image credit: Sofradir)
CSM (Calibration and Shutter Mechanism): In MSI, the two functions of calibration and shutter are gathered in one single mechanism to reduce mass, cost and quantity of mechanisms of the instrument, increasing its reliability at the same time. The CSM is located at the entrance of MSI, a rectangular device of ~ 80 cm x 30 cm, mounted on the frame of the secondary structure. The design and development of the CSM is provided by Sener Ingenieria y Sistemas, S.A., Spain. 27)
Figure 21: Photo of the CSM (Calibration and Shutter Mechanism) mechanical configuration (image credit: Sener)
Requirements and design drivers:
• During launch the CSM has to protect the instrument from sun illumination and contamination by covering the instrument entrance with a rectangular plate (named the door). This is the close position, which has to be maintained under the action of the launch loads.
• Once in orbit, the following functions are required from the CSM:
- To allow Earth observation to the instrument (MSI) the door needs to rotate from the close position 63º inwards the instrument and maintain it stable without power. This is the open position.
- From time to time, in calibration mode of the MSI, the CSM inserts a sun diffuser in front of the primary mirror and the sun diffuser is illuminated by direct solar flux. This mode corresponds to a door position located 55º from the close position outward the instrument. This position must be also stable without any power supply.
- In case of emergency, the CSM has to rotate the door to the close position from any initial position to prevent the sun light to heat sensible components of the instrument. Similarly to the previous positions, the close position shall be stable without power supply.
Figure 22: View of the CSM in calibration position (image credit: Sener)
A face to face ball bearing as rotation axis hinge in the opposite side of the actuator is used supported by means of an axially flexible support. Apart from that the pinpuller mounted on a flexible support, holds the door during launch by means of a cylindrical contact with respect to the door bushing. This design is the result of the optimization made in order to reach a stiff and robust but light and hyper-statically low constrained mechanism to make it compatible under possible thermal environments.
The pinpuller provides a reliable launch locking device and allows after pin retraction the mechanism to rotate in both senses.
The MSI instrument design represents state-of-the-art technology on many levels that is being introduced for next generation European land-surface imagers. Obviously, its performance will set new standards for future spaceborne multispectral imagers.
Storage technology introduction:
MMFU (Mass Memory and Formatting Unit):
The introduction of MMFU by EADS Astrium GmbH and IDA (Institut für Datentechnik und Kommunikationsnetze) at TU Braunschweig represents a new spaceborne storage technology based on SLC (Single Level Cell) NAND-Flash memory devices.
Note: NAND (Not And) is a Boolean logic operation that is true if any single input is false. Two-input NAND gates are often used as the sole logic element on gate array chips, because all Boolean operations can be created from NAND gates.
The NAND storage technology is not only an established technology in commercial applications but represents also a real and effective alternative for mass memory systems in space. The main advantages of the NAND-Flash technology are: a) the non-volatile data storage capability and b) the substantially higher storage density.
In the commercial world the NAND technology has become the preferred solution for storing larger quantities of data on devices such as SSDs (Solid State Drives), USB (Universal Serial Bus) Flash memory sticks, digital cameras, mobile phones and MP3-Players. In the space business, this technology has been used in some experiments only, but not in the frame of large scale mass memory systems. This is now going to be changed. 28) 29) 30)
Astrium and IDA have continuously worked for over seven years on the subject “NAND-Flash Technology for Space”. In the frame of an ESA study dubbed SGDR (Safe Guard Data Recorder) this NAND-Flash technology has been introduced and intensively evaluated.
As a result of this extensive testing, the radiation effects of this technology are well known meanwhile and appropriate error handling mechanisms to cope with the observed effects have been developed. For the S2 (Sentinel-2) mission, a complete qualification program has been performed including radiation tests, assembly qualification, construction analysis, electrical characterization, reliability tests like burn-in, destructive physical analysis, stress and life tests.
All these investments led to the final conclusion that the selected SLC NAND-Flash is an adequate technology for high capacity memory systems for space, even for systems with very high data integrity requirements.
Table 8 lists some main requirements and provides in parallel the related figures of two Astrium MMFU implementations. The first implementation is based on SLC NAND-Flash devices and will be launched with the Sentinel 2 satellite. The second option uses SDR-SDRAM devices, which was the initially required baseline technology for this mission.
Table 8: Sentinel-2 MMFU requirements and resulting implementations
The related simplified architectural block diagram of the Astrium Sentinel-2 MMFU is shown in Figure 23. The MMFU receives two parallel data streams either from the nominal or redundant VCU (Video Compression Unit). The interfaces are cross-strapped with redundant PDICs (Payload Data Interface Controllers). After reception and adaptation to internal data formats of the received source packets, the data is stored in memory modules. FMM (Flash Memory Module) and respectively SMM for the SDR-SDRAM memory module. For replay, the data is read out from two parallel operated memory modules and routed via two active TFGs (Transfer Frame Generators) providing interfaces for downlink and test. The system is controlled by a Memory System Supervisor, which is based on an ERC32 processor. The required supply voltages are provided by a power converter.
Figure 23: Architecture of the MMFU system (image credit: Astrium)
Each function is implemented by nominal and redundant hardware components. The functions and boards are summarized in Table 9:
Table 9: Number of functions and boards
Storage capacity: Astrium uses for all boards a standard format. Therefore the maximum number of memory and other devices which can be assembled on one board is limited by this form factor. Both types of memory modules are nearly identical in form, fit and function and because they can be mutually replaced; this represents a good basis for comparison.
The selected NAND-Flash device provides a capacity of 32 Gbit plus some spare. It is realized by means of four 8 Gbit dies encapsulated in a standard TSOP1 package. In total, the FMM (Flash Memory Module) includes 76 devices. The devices are arranged in four partitions which can be independently powered. A partition represents also the lowest level for reconfiguration. Each partition contains sixteen devices to store user data and three devices that are used to store parity information. This configuration enables single symbol error correction and double symbol error detection.
The SDRAM based memory module has a similar organization. There are also four partitions and each devices for single symbol error correction. A device is represented by a stack which contains eight SDRAM chips with a capacity of 512 Mbit each. From this follows the user storage capacity per memory module and some other parameters as listed in Table 10.
The number of FMM modules is determined by the total data rate and the operational concept, which requires the operation of two independent data streams. Therefore there are two memory modules operated in parallel. The third one is provided for redundancy.
The number of SMM modules is mainly determined by the required capacity. Also here two modules are operated in parallel and one SMM is included for reliability reasons.
Table 10: Performance characteristics of Astrium Sentinel 2 MMFU memory modules
The much higher storage density of the NAND-Flash devices (factor of 8) leads to a massive reduction in the number of required memory modules. For a mass memory system this becomes especially evident, if there is a requirement for a large user capacity as in case of the Sentinel-2 MMFU. Further positive aspects evolve with reduction of the number of modules. The complete system design from electrical and mechanical point of view is greatly relaxed.
Mass and volume: With reduction of the number of memory modules, it is obvious that directly related physical budgets like mass and volume, decline accordingly. Mass is always a critical issue for space missions which can be reduced by using NAND-Flash technology; but also the complete system design of a satellite, in terms of mass, power, thermal and other aspects, can be positively influenced by applying NAND-Flash based memory systems. In case of the Sentinel-2 MMFU, indeed 14 Kg (about 50%) can be saved.
Power: The power consumption is also reduced by more than 50% (Table 8). This is mainly caused by the number of memory modules operated in parallel. In case of Flash, there are only two active memory modules. In case of the SDRAM technology, 10 memory modules are operated in parallel: up to four modules for data access, two modules for read, two modules for write, and all other modules in data retention mode. Data retention means that the modules store user data and the SDRAM chips have to be refreshed and scrubbed for error detection and correction.
In contrast, a Flash-based memory module can be completely switched off without loss of data in the data retention mode. For a minimum, the partitions can be switched off and the power consumption of the controller part of the module is reduced due to low activity.
It is not obvious, that, in all cases, NAND-Flash consumes less power than SDR-SDRAM based systems. The power consumption depends on several factors like required storage capacity, data rates and operations. Generally it can be said, that as long as the required storage capacity determines the number of memory devices, Flash might be the better choice. If the number of memory devices is determined by the required data rate, SDRAM based systems may have a better performance from a power consumption point of view.
Data rates: Table 11 shows that SDR-SDRAM devices provide a much better performance from data rate point of view. The overall performance of a memory module depends on further characteristics like type of interfaces, memory controller performance, and maximum power consumption and others. Generally an SDRAM based memory module has advantages in terms of access speed.
Table 11: Performance characteristics of the memory devices
The lower performance of NAND-Flash is determined by three characteristics. During writing the NAND-Flash devices need to be programmed and this takes a time of about 700 µs per 4 kbyte data (one device page). Additionally the so-called blocks of a NAND-Flash device have to be erased before programming. This consumes another 2 ms per block (64 pages). Last but not least, the selected NAND-Flash devices use an eight bit interface for serial commanding, addressing and data transfer with a maximum operating frequency of 40 MHz.
This lack in performance can be mitigated by mainly two measures. The first straight forward measure is parallel operation of NAND-Flash devices. The second measure is interleaved access to several NAND-Flash devices. Interleaving uses the programming time of a NAND-Flash device to write in parallel the next device. These methods allow increasing the write access performance.
Life time and reliability: NAND-Flash devices provide a limited endurance. This is caused by an inherent wear out mechanism of the Flash memory cells which limits the number of erase and write cycles to about 105 cycles. To mitigate the endurance limitation, most Flash memory systems are equipped with an address management system, which distributes the write accesses rather uniformly over the address space. This technique is called Wear Leveling.
Furthermore the very high device capacity of NAND-Flash devices offers the opportunity to implement a physical address space, which exceeds the required logical user address space by a factor of n. This enhances the wear out limit of the logical addresses by the factor of n too. Hence there are two methods to keep the total count of write accesses to the same physical address below the wear out limit.
Radiation and error rates: In general, sensitivity of electronic devices to space radiation is a major topic and is also shortly discussed here through a comparison of NAND-Flash and SDR-SDRAM devices.
The mass memory system based on NAND-Flash shows clear advantages and fits well to the high storage capacity and moderate data rates of the Sentinel-2 application. The very high storage density of the NAND-Flash devices leads to a reduced number of memory modules with advantages in terms of power consumption, mass and volume. Furthermore this feature improves the reliability and eases the system design from mechanical and electrical points of view.
Figure 24: Photo of the EQM (Engineering Qualification Model), Sentinel-2 MMFU (image credit: Astrium)
Table 12: Parameters of the Sentinel-2 MMFU 31)
The ground segment includes the following elements:
• Flight Operations Segment (FOS): The FOS is responsible for all flight operations of the Sentinel-2 spacecraft including monitoring and control, execution of all platform activities and commanding of the payload schedules. It is based at ESOC, Darmstadt in Germany and comprises the Ground Station and Communications Network, the Flight Operations Control Centre and the General Purpose Communication Network.
• Payload Data Ground Segment (PDGS): The PDGS is responsible for payload and downlink planning, data acquisition, processing, archiving and downstream distribution of the Sentinel-2 satellite data, while contributing to the overall monitoring of the payload and platform in coordination with the FOS.
The Service Segment, geographically decentralized, will utilize the satellite data in combination with other data to deliver customized information services to the final users.
The baseline ground station network will include four core X-band ground stations for payload observation data downlink and one S-band station for Telemetry, Tracking and Control (TT&C). To a limited extent, the system can also accommodate some direct receiving local user ground stations for Near-Real Time applications.
The systematic activities of the PDGS include the coordinated planning of the mission subsystems and all processes cascading from the data acquired from the Sentinel-2 constellation, mainly:
1) The automated and recurrent planning of the satellite observations and transmission to a network of distributed X-band ground stations
2) The systematic acquisition and safeguarding of all spacecraft acquired data, and its processing into higher level products ensuring quality and timeliness targets
3) The recurrent calibration of the instrument as triggered by the quality control processes
4) The automated product circulation across PDGS distributed archives to ensure the required availability and reliability of the data towards users
5) The long-term archiving of all mission data with embedded redundancy over the mission lifetime and beyond.
Figure 25: PGDS context in Sentinel-2 system (image credit: ESA)
Figure 26: The Sentinel-2 ground segment (image credit: ESA)
Figure 27: Physical layout of the PGDS ground stations (image credit: ESA) 32)
• CGS (Core Ground Stations: Matera (Italy), Maspalomas (Spain), Svalbard (Norway), Alaska (USA).
• PAC (Processing/Archiving Center): Farnborough (UK), Madrid (Spain)
• MPC (Mission Performance Center): TBD
• PDMC (Payload Data Management Center): ESA/ESRIN, Frascati, Italy.
Table 13: Sentinel-2 level-1 and level-2 products
GMES / Sentinels EDRS system operations:
EDRS (European Data Relay Satellite) will provide a data relay service to Sentinel-1 and -2 and initially is required to support 4 Sentinels simultaneously. Each Sentinel will communicate with a geostationary EDRS satellite via an optical laser link. The EDRS GEO satellite will relay the data to the ground via a Ka-band link. Optionally, the Ka-band downlink is planned to be encrypted, e.g. in support to security relevant applications. Two EDRS geo-stationary satellites are currently planned, providing in-orbit redundancy to the Sentinels. 33)
EDRS will provide the same data at the ground station interface as is available at the input to the OCP (Optical Communications Payload) on-board the satellites, using the same interface as the X-band downlink. The EDRS transparently adapts the Sentinels data rate and format to the internal EDRS rate and formats, e.g. EDRS operates at bit rates of 600 Mbit/s and higher.
With EDRS, instrument data is directly down-linked via data relay to processing and archiving centers, while other data continues to be received at X-band ground stations. The allocation of the data to downlink via X-band or EDRS is handled as part of the Sentinel mission planning system and will take into account the visibility zones of the X-band station network and requirements such as timeliness of data.
Figure 28: Sentinel missions - EDRS interfaces (image credit: ESA)
GMES / Sentinel data policy:
The principles of the Sentinel data policy, jointly established by EC and ESA, are based on a full and open access to the data:
• anybody can access acquired Sentinel data; in particular, no difference is made between public, commercial and scientific use and in between European or non-European users (on a best effort basis, taking into consideration technical and financial constraints);
• the licenses for the Sentinel data itself are free of charge;
• the Sentinel data will be made available to the users via a "generic" online access mode, free of charge. "Generic" online access is subject to a user registration process and to the acceptation of generic terms and conditions;
• additional access modes and the delivery of additional products will be tailored to specific user needs, and therefore subject to tailored conditions;
• in the event security restrictions apply to specific Sentinel data affecting data availability or timeliness, specific operational procedures will be activated.
1) P. Martimort, O. Arino, M. Berger, R. Biasutti, B. Carnicero, U. Del Bello, V. Fernandez, F. Gascon, B. Greco, P. Silvestrin, F. Spoto, O. Sy, “Sentinel-2 Optical High Resolution Mission for GMES Operational Services,” Proceedings of IGARSS 2007 (International Geoscience and Remote Sensing Symposium), Barcelona, Spain, July 23-27, 2007
2) P. Martimort, F. Spoto, B. Koetz, O. Arino, M. Rast, “GMES Sentinel-2, The Optical High Resolution Mission for GMES Operational Services,” July 16, 2007, URL: http://www.earthobservations.org/documents/cop/ag_gams/200707_01/gmes_sentinel2.pdf
3) ESA Sentinel-2 Team, “GMES Sentinel-2 Mission Requirements Document,” Issue 2, Revision 1, 08.03.2010, URL: http://esamultimedia.esa.int/docs/GMES/Sentinel-2_MRD.pdf
4) F. Spoto, M. Berger, “Sentinel-2 Presentation,” March 7-8, 2007, ESA/ESRIN, URL: http://esamultimedia.esa.int/docs/GMES/ESA/4_coloc_GMES_S2_GSC_day_7_8_March_2007.pdf
5) P. Martimort, M. Berger, B. Carnicero, U. Del Bello, V. Fernandez, F. Gascon, P. Silvestrin, F. Spoto, Omar Sy, O. Arino, R. Biasutti, B. Greco, “Sentinel-2 : The Optical High Resolution Mission for GMES Operational Services,” ESA Bulletin, No 131, Aug. 2007, pp. 19-23, URL: http://www.esa.int/esapub/bulletin/bulletin131/bul131b_martimort.pdf
6) Philippe Martimort, “The Optical High Resolution Mission The for GMES Operational Services for Services,” GMES Sentinel-2, AGRISAR Workshop, Noordwijk, The Netherlands, Oct. 15-16, 2007, URL: http://www.dlr.de/hr/Portaldata/32/Resources/dokumente/agrisar/(03)_Presentation_AGRISAR.pdf
7) F. Spoto, P. Martimort, O. Sy, P. Bargellini, B. Greco, “Sentinel-2: the European operational fast revisit high resolution land observing mission,” Sentinel-2: the European operational fast revisit high resolution land observing mission
8) François Spoto, Philippe Martimort, Omar Sy, Paolo Laberinti, Stefane Carlier, Umberto Del Bello, Valérie Fernandez, Volker Kirschner, Claudia Isola, Matthias Drusch, Pier Bargellini, Franco Marchese, Olivier Colin, Ferran Gascon, Bianca Hoersch, Aimé Meygret, “The GMES Sentinel-2 Operational Mission,” Proceedings of the ESA Living Planet Symposium, SP-686, Bergen Norway, June 28-July 2, 2010
9) “Copernicus: new name for European Earth Observation Programme,” European Commission Press Release, Dec. 12, 2012, URL: http://europa.eu/rapid/press-release_IP-12-1345_en.htm
11) GMES Sentinel-2 Industry Day, Nov. 26, 2007, URL: http://esamultimedia.esa.int/docs/industry/Sentinel/Nov07/Sentinel-2-Industry-Day-Nov-07-Astrium-VGs.pdf
13) Francois Spoto, Omar Sy, Paolo Laberinti, Philippe Martimort, Valerie Fernandez, Olivier Colin, Bianca Hoersch, Aime Meygret, “Overview of Sentinel-2,” Proceedings of IGARSS (International Geoscience and Remote Sensing Symposium), Munich, Germany, July 22-27, 2012
14) “Astrium to build ESA's second Sentinel-2 satellite for GMES,” March 31, 2010, URL: http://www.esa.int/esaLP/SEME2GIK97G_LPgmes_0.html
15) S. Winkler, G. Wiedermann, W. Gockel, “High-Accuracy On-Board Attitude Estimation for the GMES Sentinel-2 Satellite: Concept, Design, and First Results,” AIAA Guidance, Navigation and Control Conference and Exhibit, August 18-21, 2008, Honolulu, Hawaii, USA, AIAA 2008-7482
16) “Sentinel-2,” ESA, URL: /web/guest/missions/esa-future-missions/sentinel-2
18) Robert Lange, Frank Heine, Hartmut Kämpfer, Rolf Meyer, "High Data Rate Optical Inter-Satellite Links," 35th ECOC (European Conference on Optical Communication) Sept. 20-24, 2009, Vienna, Austria
20) Vincent Cazaubiel, Vincent Chorvalli, Christophe Miesch, “The Multispectral Instrument of the Sentinel-2 Program,” Proceedings of the 7th ICSO (International Conference on Space Optics) 2008, Toulouse, France, Oct. 14-17, 2008
21) Michel Bréart de Boisanger, Olivier Saint-Pé, Franck Larnaudie, Saiprasad Guiry, Pierre Magnan, Philippe Martin Gonthier, Franck Corbiere, Nicolas Huger, Neil Guyatt, “COBRA, a CMOS Space Qualified Detector Family Covering the Need for many LEO and GEO Optical Instruments,” Proceedings of the 7th ICSO (International Conference on Space Optics) 2008, Toulouse, France, Oct. 14-17, 2008
22) François Spoto , Philippe Martimort, Omar Sy, Paolo Laberinti, “Sentinel-2, Optical High Resolution Mission for GMES Operational services,” Sentinel-2 Preparatory Symposium, ESA/ESRIN, Frascati, Italy, April 23-27, 2012, URL: http://www.s2symposium.org/
23) Vincent Chorvalli, Stéphane Espuche, Francis Delbru, Cornelius Haas, Philippe Martimort, Valérie Fernandez, Volker Kirchner, “The Multispectral Instrument of the Sentinel-2 Em Program Results,” Proceedings of the ICSO (International Conference on Space Optics), Ajaccio, Corse, France, Oct. 9-12, 2012, paper: ICSO-023
24) Michel Bougoin, Jerome Lavenac, “The SiC hardware of the Sentinel-2 Multi Spectral Instrument,” Proceedings of the ICSO (International Conference on Space Optics), Ajaccio, Corse, France, Oct. 9-12, 2012, paper: ICSO-028
25) P. Gloesener, F. Wolfs, F. Lemagne, C. Flebus, “Manufacturing, testing and alignment of Sentinel-2 MSI telescope mirrors,” Proceedings of the ICSO (International Conference on Space Optics), Ajaccio, Corse, France, Oct. 9-12, 2012 , paper: ICSO-034
26) Karin Schröter, Uwe Schallenberg, Matthias Mohaupt, “Technological Development of Spectral Filters for Sentinel-2,” Proceedings of the 7th ICSO (International Conference on Space Optics) 2008, Toulouse, France, Oct. 14-17, 2008
27) J. A. Andion, X. Olaskoaga, “Sentinel-2 Multispectral Instrument Calibration and Shutter Mechanism,” Proceedings of the 14th European Space Mechanisms & Tribology Symposium – ESMATS 2011, Constance, Germany, Sept. 28–30 2011 (ESA SP-698)
28) M. Staehle, M. Cassel, U. Lonsdorfer l, F. Gliem, D. Walter, T. Fichna, “Sentinel 2 MMFU: The first European Mass Memory System Based on NAND-Flash Storage Technology,” Proceedings of the DASIA (DAta Systems In Aerospace) 2011 Conference, San Anton, Malta, May 17-20, 2011, ESA SP-694, August 2011
29) M. Staehle, M. Cassel, U. Lonsdorfer, F. Gliem, D. Walter, T. Fichna, “Sentinel-2 MMFU: The first European Mass Memory System based on NAND-Flash Storage Technology,” Proceedings of ReSpace/MAPLD 2011, Aug. 22-25, 2011, Albuquerque, NM, USA, URL: https://nepp.nasa.gov/respace_mapld11/talks/thu/ReSpace_C/1030%20-%20Cassel.pdf
30) Giuseppe Mandorlo, “Sentinel-2 Mass Memory and Formatting Unit and Future File Based Operations,” Proceedings of ADCSS (Avionics Data, Control and Software Systems) Workshop, ESA/ESTEC, Noordwijk, The Netherlands, Oct.23-25, 2012, URL: http://congrexprojects.com/docs/12c25_2510/06mandorlo_mmfufileops.pdf?sfvrsn=2
31) Michael Stähle, Tim Pike, “ADCSS 2012 Astrium - Current and Future Mass Memory Products,” Proceedings of ADCSS (Avionics Data, Control and Software Systems) Workshop, ESA/ESTEC, Noordwijk, The Netherlands, Oct.23-25, 2012, URL: http://congrexprojects.com/docs/12c25_2510/09stahele_astriumfinal.pdf?sfvrsn=2
33) H. L. Moeller, S. Lokas, O. Sy, B. Seitz, P. Bargellini, “The GMES-Sentinels – System and Operations,” Proceedings of the SpaceOps 2010 Conference, Huntsville, ALA, USA, April 25-30, 2010, paper: AIAA 2010-2189
34) Henri Laur, “SAR Interferometry opportunities with the European Space Agency: ERS-1, ERS-2, Envisat, Sentinel-1A, Sentinel-1B, ESA 3rd Party Missions (ALOS),” Fringe 2009 Workshop - Advances in the Science and Applications of SAR Interferometry, Frascati, Italy, Nov. 30-Dec. 4, 2009
35) “ESA Member States approve full and open Sentinel data policy principles,” ESA, Nov. 27, 2009, URL: http://www.esa.int/esaEO/SEMXK570A2G_environment_0.html
36) Susanne Mecklenburg, “GMES Sentinel Data Policy - An overview,” GENESI-DR (Ground European Network for Earth Science Interoperations - Digital Repositories) workshop, ESAC, Villafranca, Spain, December 4, 2009
37) Bianca Hoersch, “GMES Space Component & Sentinel(-2),” Landsat Science Team Meeting, Mountain View, CA, USA, Jan. 19-21, 2010, URL: http://landsat.usgs.gov/documents/Jan_2010_Landsat_Science_Team_meeting_Jan2010_Hoersch_Final-short.pdf
The information compiled and edited in this article was provided by Herbert J. Kramer from his documentation of: ”Observation of the Earth and Its Environment: Survey of Missions and Sensors” (Springer Verlag) as well as many other sources after the publication of the 4th edition in 2002. - Comments and corrections to this article are always welcome for further updates.
The Sentinel series:
Provides data continuity for: