THE SUSTAINABILITY OF THE RAPIDEYE REMOTE SENSING CONSTELLATION

Transcription

1 THE SUSTAINABILITY OF THE RAPIDEYE REMOTE SENSING CONSTELLATION Enrico Stoll (1), Kam Shahid (1), Erika Paasche (1), Marcus Apel (1) (1) BlackBridge, Kurfürstendamm 22, Berlin, Germany; {-415, -405, -303, -510}; {enrico.stoll, kam.shahid, erika.paasche, ABSTRACT BlackBridge is focused on providing end to end solutions across the geospatial value chain. The resource for the geospatial information services is the RapidEye satellite constellation consisting of five satellites, owned and operated by BlackBridge. The orbital distribution of the spacecraft allows for frequent revisits and the acquisition of more than five million square kilometres of multispectral imagery per day. In order to steadily increase the degree of utilization capacity of the satellites, the mission planning and operations are dynamic entities that are subject to modifications and optimizations. Feasibility analyses are performed before long-term imaging campaigns are initiated to assure that the area of interest can be covered efficiently under complex boundaries, such as cloud distributions and seasonal influences. This paper elucidates the efforts that are being undertaken to guarantee the sustainability of the RapidEye remote sensing constellation. It correlates the theoretical feasibility analyses with data from representative imaging campaigns. Flight results, obtained during a recent subsystems performance review for the entire constellation are also highlighted. 1 INTRODUCTION BlackBridge owns and operates an end-to-end commercial Earth observation system, comprised of the RapidEye constellation of 5 spacecraft, a dedicated spacecraft control center (SCC), reception facilities for enabling uplink/downlink services, and a ground segment designed to plan, acquire, and process over 5 million km² of imagery every day. Five years of operational service have proven BlackBridge's capability to react fast and efficiently to varying customer requests. Built by Surrey Satellite Technology Ltd. (SSTL) in England, each satellite is based on an evolution of the flight proven SSTL-150 platform. Each of the RapidEye satellite measures less than one cubic meter and weighs 150 kg. Onboard digital recorders store image data until each satellite passes within range of ground receiving stations located in Svalbard, Norway. The main constellation properties are summarized in Table 1. Table 1: Satellite Platform Performance Property Value Spacecraft Roll Angle +/-25 degrees Global Revisit Time (All) 1 day Pointing Control < 10 m (2σ) Inclination 97.8 Orbit Sun Synchronous Period 96.7 min Descending Node 11 a.m. (approximately) The 4S Symposium 2014 E. Stoll 1

2 The sun synchronous orbit in combination with the satellite spacing on the orbit allow for a daily revisit capability. The multispectral imager was developed by Jena-Optronik, utilizing a Three Mirror Anastigmatic (TMA) design and permitting a Field of View of 7. This results in a swath width of 77 km which is complemented by the capability of the system to acquire image takes of up to 1500 km per orbit. Figure 1 shows an example of image acquisitions of the constellation during one day, one week, and one month. Figure 1: Sample image acquisition during one day (left), one week (middle), and one month (right) 2 SATELLITE TASKING AND FEASIBILITY ANALYSES BlackBridge coordinates various imaging campaigns to efficiently task the RapidEye constellation. The prerequisite for an optimal utilization of the system are feasibility analyzes as will be explained in this section. 2.1 Satellite Tasking The current satellite tasking is based on the RapidEye tile grid and the individual tile score [2]. The entire world is separated into a custom system of approx. 25 km x 25 km tiles. Figure 2 shows an example of Berlin. Depending on the weather forecast and the importance of an individual tile (e.g. due to customer orders), a tile score can be calculated for each tile. The satellite tasking algorithm subsequently considers all tiles along the ground track of a RapidEye satellite and selects the roll angle that maximizes the sum of the tile score along the path. In order to incorporate the latest weather forecast, acquisition planning for the constellation is done twice a day. The morning planning session plans for image acquisitions in North and South America (acquired between 13:00 24:00 UTC) and the evening planning session sets up imaging for the orbits over Australia, Asia, Europe, and Africa (acquired between 0:00 13:00 UTC). The finalized plans are then uploaded through RapidEye s spacecraft control centre (SCC) to schedule the satellites. The 4S Symposium 2014 E. Stoll 2

3 Figure 2: Tiles associated with Berlin After establishing a long term service contract with Kongsberg Satellite Services AS (KSAT) BlackBridge has the possibility of downloading the imagery data to an X-band ground station in Svalbard, Norway providing the necessary bandwidth as well as an advantageous position at high latitude. Thus, every orbit of the RapidEye satellites can be supported by KSAT. Besides imagery, also received Payload Ancillary Files (PAF) and Bus Ancillary Files (BAF) are forwarded by KSAT using terrestrial communication lines to Berlin headquarters. In order to prepare the receipt of these files from the constellation, KSAT evaluates the reception schedules, provided by RapidEye, and also containing the latest two-line elements (TLE) of the satellites. 2.2 Feasibility Analyses Before large imaging campaigns are introduced into the planning system, a feasibility analyses are carried out in order to evaluate the time necessary to image a given Area of Interest (AOI). Such an analysis incorporates the satellite orbits and appropriate orbit propagators. Statistical cloud data, which is available as monthly mean cloud coverage over 25 years, is used to simulate cloud distribution during the time of interest (TOI). The statistical cloud data is available per latitudelongitude position as Figure 3 shows. Figure 3: Statistical cloud coverage [%] for June The feasibility analysis tool subsequently evaluates the area that can be covered at a given time and accumulates the image takes that feature a cloud cover that is less than a to be specified value (20 % is default). Figure 4 shows the result of an example coverage of Germany in June. The graph on the right shows the coverage progression by time, whereas the areal representation shows the covered The 4S Symposium 2014 E. Stoll 3

4 as well as the remaining AOI after the simulation is finished (after 30days). Figure 4: Theoretical coverage of Germany in June Figure 5 shows acquired tiles during an imaging campaign in June It emphasizes that the coverage predictions are usually in good accordance with reality and thus, appropriate to evaluate the tasking strategy for future imaging campaigns. It is used to specify AOI - TOI combinations and resolve potential competitions between customer orders. Figure 5: Acquisition over Germany in June 2010 The 4S Symposium 2014 E. Stoll 4

5 3 THE SUSTAINABILITY OF THE RAPIDEYE CONSTELLATION This section highlights several efforts that are being undertaken to guarantee the sustainability of the remote sensing constellation. 3.1 Orbit Maintenance RapidEye satellites are subject to perturbing accelerations, which result in a continuous altitude decrease of the constellation. Orbit maintenance manoeuvres are necessary to keep the constellation in an optimal formation with respect to the imaging and daily revisit capabilities. These maintenance operations ensure that the satellites are located at a reference orbit of around 630 km altitude [3]. The orbital elements are further optimized for a target phase difference of approx. 72 between the satellites. This assures both the daily revisit capabilities of the constellation and the successive downlink possibility at one ground station. Figure 6 shows the drop rates of the five satellites since the beginning of the mission. The data is derived from the on-board GPS. While the drop rates where comparably low during the first two years (around 1 m/s) they show a considerable increase in magnitude thereafter. Drop rates of up to 12m/s necessitate an increase of the manoeuvre size to compensate for the decay due to solar activities. Typical manoeuvres in 2013 comprised height changes of about 400 m and velocity changes of about 0.2 m/s. Figure 6: Drop rates of the five RapidEye satellites 3.2 Collision Probability Assessment The number of satellites in Earth orbit is steadily growing and with the high amount of space debris, either crossing through or resident in orbit, collision probabilities between two such objects can become critical. 126 close approaches between the RapidEye constellation and secondary space objects were detected as of March A summary is given in Table 2. These close approaches constitute an increased risk for RapidEye's remote sensing satellites not only during nominal operations but also during orbit maintenance periods. BlackBridge has taken the approach to partner with external agencies to provide assessments of collision probability [4]. The European Space Agency (ESA) was chosen in this capacity, since the Space Debris Office has a long history in space debris research, operational collision risk analysis and avoidance manoeuvre planning. BlackBridge provides constellation GPS data to the ESA SDO on a daily basis. The collision probability threshold, below which a manoeuvre is initiated is As Figure 7 shows, this value is associated with 0.2 collision avoidance manoeuvres per year and satellite [5]. Thus, 1 overall collision avoidance manoeuvre per year is anticipated which is in accordance with the Eye s recent manoeuvre history of the last two years. The 4S Symposium 2014 E. Stoll 5

6 Table 2: 126 conjunctions as of March Number of Events per Year - scaled with factor 2 RE events examples 1 22 SL-14 R/B (10x) Cosmos 2251 DEB (4x) 2 18 Fengyun 1C DEB (3x) Pegasus R/B (2x) 3 38 SL-14 R/B (8x) Fengyun 1C DEB (5x) 4 29 Fengyun 1C DEB (6x) SL-14 R/B (4x) 5 19 Cosmos 2251 DEB (5x) Fengyun 1C DEB (3x) Num ber of Events per Year [-] e-006 1e Accepted Collision Probability Level [-] Figure 7: Manoeuvre rate as a function of the accepted collision probability level [5] In order to have minimal impact on the data acquisition process, the decision to introduce a collision avoidance manoeuvre must be made before the respective planning session starts. For manoeuvres during the orbits that image over North and South America a decision must be taken by 7:00 UTC on the same day, whereas for the remaining regions, the decision must be taken by 12:30 UTC on the previous day. This poses an additional boundary condition on the collision avoidance manoeuvre planning 3.3 Subsystems Performance The design life is the period of time during which a system is expected by its designers to work within its specified parameters. The design lifetime of each of the satellites within the RapidEye constellation is 7.25 years following launch. It was determined pre-launch by the consumption rate of relevant consumables as well as by time dependent functionalities within distinct subsystems. Within the five years of operations all subsystems were closely monitored to predict the actual operational life-time of the satellites. A summary of the latest investigations is given by subsystem. It shows that, given the current status, and operating criteria of the RapidEye constellation, the predicted mission operation lifetime will outperform the design life by several years. Propellant The consumption rate of the propellant is and has always been considerably lower than assumed during the design phase. Very little propellant was used to correct for launch inaccuracies. Furthermore the impact of environmental conditions, in particular solar activity, have shown to be much lower than predicted. The atmospheric drag experienced by a satellite is lower in the case of reduced solar activity, and vice versa. Instead of predicted drop rates of 20 m / day the satellites experienced average drop rates of 5 m /day (cp. Figure 6). Thus, the need to use propellant to lift the satellites is considerably lower. The current fuel level along with the predicted fuel use based on the current orbit maneuvering strategy shows that the propellant will not be a limiting factor for the constellation within the next 30 years. Battery The ability of a battery to repetitively store energy up to a required amount fades over time. Assessing telemetry data over the last five years has allowed the refinement of the battery fade model to better predicted the fade behavior. Thus, the predicted lifetime of the battery can be approximated by extrapolating the measured performance curve, which is assumed to be linear. It shows that the batteries will still support nominal operations after The 4S Symposium 2014 E. Stoll 6

7 Solar cells The ability of solar cells to convert solar light into electrical energy is known to fade over time. To allow a nominal operation of the satellites a distinct amount of energy has to be generated every day. Currently, the solar cells generate enough power for both imaging operations and to charge the batteries to their full capacity. Also, there is additional margin above our current needs for charging capability. Camera Sensors Up to this point in time there are no stuck or dead pixels in any of the RapidEye payloads. Special ground based algorithmic approaches are available to mitigate the risk of limited stuck and dead pixel occurrences. The common approach to maintain a stable response of all Charge-Coupled Device (CCD) sensors over time is to apply calibration tables. Significant efforts are being undertaken to update these calibration tables according to the actual sensitivity of each CCD by applying statistical methods but also by conducting regular in-situ measurements on the ground. Mechanical The reaction wheels are the primary actuators for the attitude control system. All satellites can perform full roll maneuvers (+20 to -20 degrees) in order to maintain the constellation daily revisit requirement. 4 Further System Optimizations The optimization of the RapidEye system is an ongoing process. This section highlights some processes which were introduced in that framework. 4.1 Orbit Position Determination When RapidEye started operations, orbit maintenance maneuvers usually affected the operations for seven further days. The reason for this was not the image acquisition itself but the data downlink over Svalbard. RapidEye was relying on the two-line element (TLE) data of the North American Aerospace Defense Command (NORAD) for forwarding information on the spacecraft position to KSAT. KSAT needs the information for tracking the orbital positions of the satellites and maintain communication with the spacecraft. Figure 8: Comparison of SMA data by NORAD and by on-board GPS after orbital manoeuvres Since NORAD s TLE predictions are based on ranging results over a time span of multiple days, data points of before and after the burn are taken into consideration to generate the spacecraft TLE. The 4S Symposium 2014 E. Stoll 7

8 Thus, the state vector predictions were not accurate enough to support the tracking of the spacecraft with modified altitude. Reception failures and the loss of image data were the consequence. Figure 8 shows an example of RapidEye-2 and RapidEye-5. The semi-major axis (SMA) obtained through the use of NORAD TLE and determined by on-board GPS are compared. It shows that a difference of up to 300m can occur and it takes approx. 7 days for the two curves to converge. Subsequently, BlackBridge developed a method to generate their own TLEs based on the spacecraft GPS after the burn. TLEs were sent to KSAT and proved to work efficiently during orbit maintenance periods. 4.2 Calibration Techniques Ideally every detector in the focal plane of a remote sensing system would have the same response (mean and standard deviation) when stimulated by the same input light signal. Thus, an image of a uniform radiance input scene would itself be uniform and free from any variations in brightness. However, this is rarely the case and the raw due to fixed pattern noise. These spatial intensity variations can be attributed to differences in individual detector photo-response, electronic signal processing, throughput variation as a function of the field-of-view of the optical system, and the presence of contaminants within the payload environment. Typically detector responses also change over time during the mission. In an effort to improve the relative spatial calibration (i.e. detector response equalization) of the MSI a side-slither calibration method was developed [7]. Pseudo-invariant and spatially uniform terrestrial scenes that included desert and snow/ice fields were imaged with the sensor in a ninetydegree yaw orbital configuration. In this configuration, each detector on the focal plane was positioned parallel to the ground-track direction thereby imaging the same segment of ground and exposing each detector to the same target radiance. This manoeuvre produced a radiometrically flatfield input to the sensor so that the relative response of each detector was determined for the same exposure level and compared to the array average. While roll maneuvers along the in-track direction are part of the daily imaging activities, such ninety-degree yaw maneuvers in connection with an imaging activity were never foreseen by the manufacturer. Thus, extensive test campaigns, in which e.g. battery and thermal conditions of the spacecraft were initially simulated and then monitored on orbit, had to be performed before the yaw procedure and the side-slither method were incorporated into daily operations [6]. 4.3 Direct Downlink The original design of the RapidEye system was based on a central image date downlink station. From there the image data is re-distributed. On demand of users, RapidEye extended their system to allow direct reception of image data [1]. This RapidEye Direct DownLink (DDL) service provides users the capability to directly receive image data from the satellite constellation to their own system. The receiving antenna location may be located anywhere in the world between ±84 latitude. RapidEye offers DDL users the direct reception of data from overhead passes and preceding orbits. Preceding orbit image takes take advantage of the satellites capabilities to carry their on-board data storage for one orbit until a suitable telemetry downlink contact is available. The possible Area of Interest (AOI) for the DDL users is therefore extended significantly. The data transmitted to the DDL customer follows the European Space Agency (ESA) implementation of the CCSDS standards. The spacecraft transmit in X-Band three different file types, namely Level 0 image data band files, PAFs, and BAFs. The ordering of data can be done in several levels of urgency: Nominal tasking for a timespan to be defined by the DDL user. Repetitive tasking over several successive time spans. Rapid or emergency tasking where the user intends to get immediate data over a specified The 4S Symposium 2014 E. Stoll 8

9 area irrespective of the actual cloud cover situation. 5 SUMMARY BlackBridge provides highly reliable data products using its satellite constellation. Serious efforts are undertaken to adapt to a changing market and boundary conditions. This paper outlined the capabilities of the five RapidEye spacecraft and an ongoing optimization process. It showed the end-to-end system including image planning, tasking, and feasibility analyses. It further highlighted system enhancements, which had to be performed to optimize operations. In this connection, a direct downlink capability was developed to deliver near real-time data, a new calibration technique is implemented into daily operations, and a collision avoidance capability was developed. It is shown that the RapidEye constellation can provide high quality data for large AOIs with a high temporal resolution. 6 REFERENCES [1] H. Konstanski, B. D Souza, E. Stoll, M. Oxfort, Four years of operating RapidEye satellites: Continuous improvement of performance and adaptation to customer markets, IAA Symposium on Small Satellites for Earth Observation, Berlin, Germany, [2] E. Stoll, H. Konstanski, C. Anderson, K. Douglass, M. Oxfort, The RapidEye Constellation and its Data Products, proc. of IEEE Aerospace Conference, Big Sky, USA, [3] E. Stoll, R. Schulze, B. D Souza, M. Oxfort, The impact of collision avoidance maneuvres on satellite constellation management, proc. of European Space Surveillance Conference, Madrid, Spain, 2011 [4] E. Stoll, B. B. Virgili, K. Merz, H. Krag, B. D Souza, Collision Probability Assessment for the RapidEye Satellite Constellation, European Conference on Space Debris, Darmstadt, Germany, [5] E. Stoll, B. B. Virgili, K. Merz, H. Krag, B. D Souza, Operational Collision Avoidance of Small Satellite Missions, IEEE Aerospace Conference, Big Sky, USA, [6] Godard, E. Stoll, C. Anderson, R. Schulze, B. D Souza, Integrating Advanced Calibration Techniques into Routine Spacecraft Operations, proc. of SpaceOps 2012, Stockholm, Sweden, Jun [7] C. Anderson, D. Naughton, A. Brunn, M. Thiele, Radiometric correction of RapidEye imagery using the on-orbit side-slither method, in proc. of SPIE Remote Sensing, Prague, Czech Republic, The 4S Symposium 2014 E. Stoll 9