Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Distributed Ground Station Network @ IAC-13 (paper)


Published on

Published in: Business, Technology
  • Login to see the comments

Distributed Ground Station Network @ IAC-13 (paper)

  1. 1. 64th International Astronautical Congress, Beijing, China. This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License, 2013. IAC-13,B4,3,11,x17101 DISTRIBUTED GROUND STATION NETWORK - A GLOBAL SYSTEM FOR TRACKING AND COMMUNICATION WITH SMALL SATELLITES AS AN OPEN SERVICE Mr. Andreas Hornig University of Stuttgart, Germany, Mr. Timm Eversmeyer Germany, Mr. Ulrich Beyermann University of Stuttgart, Germany, Small satellite missions face two special challenges due to limited financial budgets. The first is tracking the satellite for orbit determination after orbit injection. The orbit can vary from the specified orbit when the satellite will be launched as secondary or parasitic payload. In case of unknown orbit parameters high-gain antennas cannot be pointed towards the satellite and establish the first connection needed for satellite activation. The second challenge is to transmit all housekeeping and scientific data to mission operations via a limited number of ground-stations. The Distributed Ground Station Network (DGSN) solution can solve the problem with permanent tracking and a faster orbital element provision to the satellite owners. And it can provide permanent reception of satellite signals with its data-dump mode in between main ground-stations provided by the satellite owner. The key feature is the network of small ground-stations placed globally connected via the internet and performing an automatic scan of satellite (and other beacon) signals, storing and sending them back to a central server, where they can be accessed by the satellite owner. With a correlation of the beacon signal and GNSS synched ground-station time the satellite position is determined with pseudo-ranging trilateration. In contrast to ground-station time sharing concepts of Radio Aurora Explorer (SRI International in California and University of Michigan), AISat (DLR) and QB50-GENSO (Karman Institute for Fluid Dynamics) that rely on a limited number of amateur radio operators and expensive hardware with limited availability DGSN uses an innovative citizen science approach. The participatory aspect includes the deployment of ground-station hardware and acquisition of satellite signal data but also the open-source hardware. In this way a high number of built sensornodes can be achieved and it also creates a new market for selling ready ground-stations. The network owners an open-platform for every small satellite operator with a faster access to tracking data than the update period by NORAD or ESTRACK with less running costs. The low data-rate reception is compensated by the global and permanent coverage. DGSN will offer the orbital parameters of the received satellite signal under a free license. The feasibility study had been conducted as part of Azorean observing VERDE Sat during the Small Satellite Project at the Institute of Space Systems (IRS) of the University of Stuttgart and DGSN is in the prototyping and testing phase of the ground-stations. I. INTRODUCTION For small satellites that will be launched as secondary or parasitic payloads the injected orbit can vary from the specified one. This is due to the fact that the main payload launched with the same launching system sets the main target orbit and the launching system is programmed to achieve this orbit with high accuracy and the parasitic payload's orbit can be influenced and thus be slightly changed. This change in orbit has to be known by the small satellite owner to be able to establish a communication link. When the orbital elements differ high gain antennas cannot be pointed to the satellite and miss the target's on-board antenna. Therefor a relatively fast tracking and positioning of the satellite is required. Setting up an own tracking system with multi ground stations is costly and most often not suitable for IAC-13-B4,3,11,x17101 small satellite owners like universities. The utilization of using external databases of satellite ephemerides by NORAD and ESTRACK is a challenge for the satellite mission planning because the databases are updated on a regular basis sometimes not matching the launch time. A second aspect is the network of possible ground stations that can be used for telemetry and control of such small satellites. A permanent coverage of satellite signal reception is costly or requires partnerships with commercial stations or with the amateur radio community. II. PROPOSED NETWORK SOLUTION The financial and planning tasks are one of the main challenges for the VERDE sat (VEgetation Research and DEtection) during the Small Satellite Project [1] at the Institute of Space Systems (IRS) of the University of Page 1 of 11
  2. 2. 64th International Astronautical Congress, Beijing, China. This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License, 2013. Stuttgart. The mission objective is to observe the indigenous and exotic vegetation on the Azore islands with a small satellite. This satellite budget is 5 M€, it is booked for a launch with the Indian GSLV launcher system and it has access to two primary ground stations at the IRS and on the campus of the University of the Azores, São Miguel. The secondary launch option into a sun synchronous orbit (575 km height) with an uncertainty of the provided orbital element in combination with the high-gain antennas and limited access times to the primary ground stations resulted in the concept of the DGSN. It was designed to provide a simple and cost efficient tracking system with respect to on-board satellite hardware as well as the ground segment. The allocation of amateur radio bands for the highspeed payload data and the TT&C channels also requires including the ham-radio community into satellite mission to be allowed to use the bands. With low-cost ground station hardware not only the hamradio community but also the participatory ordinary citizen can be part in a grass-roots space program. In this way a huge basis of only-receiving ground stations can be deployed forming a global grid of distributed ground stations via the internet. This network can be used for tracking satellite signals, it gives the opportunity of globally collecting satellite housekeeping and even payload data, and it also includes the public and brings space research to the front gardens as well as balconies of the everyman. So it offers a synergy between the satellite owner and providing a fast and permanent communication and tracking service and it directly involves the citizen with an active and important part within a satellite mission. The implementation of the beacon hardware onboard of the satellite is minimal, because it uses the existing communication payload so the DGSN system was chosen as the mandatory secondary science payload of VERDE. Additional missions like the FlyingLaptop, the flagship mission of the Small Satellite Project at the IRS [1], the ArduSat [2] cube sat by nanosatisfi and the FREDE [3] experiment, on-board the high-altitude balloon mission BEXUS, applied for candidates. As one result of the study the DGS-network is designed as an open service for further satellites using the network. The concept even proceeds existing concepts by including everyone and not only the professional or ham-radio community, where ham licenses are required. The participation as a node of the reception grid is also simplified by an open-source approach of the ground station hardware. So interested users can either buy a starter kit or build their own station. This also allows a creative utilization of the concept and spawns further sensor ideas by adding modules for further tasks like GPS signal measuring similar to EGNOS, adding further antennas and thus IAC-13-B4,3,11,x17101 frequency ranges or completely new applications due to lateral thinking. III. DISTRIBUTED GROUND STATION NETWORK The distributed ground station network (DGSN) solution can solve the problem with permanent tracking and a faster orbital element provision to the satellite owners. And it can provide permanent reception of satellite signals with its data-dump mode in between main ground stations provided by the satellite owner. In this way it reduces black-out times and provides the advantage of a low data-rate, but constant and permanent reception globally. This is essential for small projects with limited budget and the DGSN can provide a service these parties do not have to build and operate by themselves and it can provide some kind of standard for an open use. III.I DGSN Architecture The DGSN architecture is divided in space, ground and network segments (Fig. 1). Fig. 1: Tracking and communication sequence (a-d) The key feature is the network of small ground stations placed globally connected via the internet and performing an automatic scan for satellite (and other beacon) signals, storing and sending them back to a central server, where they can be accessed by the satellite owner. This is part of the ground segment. The network segment, formed by ground stations on multiple locations and connected via the Internet, makes Page 2 of 11
  3. 3. 64th International Astronautical Congress, Beijing, China. This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License, 2013. it possible to track them with trilateration, but it needs a precise system time to correlate the received signals with the reception time. That can be done with global navigation satellite systems (GNSS) like GPS, Galileo and others, So each ground station will be equipped with a GNSS receiver that also determines the stations' locations. The ground station will be connected to a standard personal computer with internet access. An open-source software called BOINC (Berkeley Open Interface for Network Computing) will be used to attach the PC to the Constellation Platform [4], from where the communication application will be loaded. Afterwards the PC will automatically perform the needed signal scanning with the connected ground station hardware. The received data will be sent to the central Constellation server where it will processed. And another feature is signal processing and processing other numerical tasks on the PCs connected to ground stations. In this way also computing capacity for a distributed computing system is provided. The Constellation Platform is already available for this application. The ground station coverage relies on communities ranging from the owner itself, amateur radio operators and volunteers, for this scenario the ground station hardware has to be simple and cheap, and the first ground station version will be used for passive signal reception. That means that it will only receive signals for tracking and data dumping, but not for signal transmission from ground station to the satellite. This is due to operator licensing and because of safety reasons for commanding signals. So the DGSN is an expansion of the current communication and tracking system and offers a niche especially for small satellites. The network segment is also designed to work without a central server in the internet in a 5+ ground station mode and only connected on a local area or even without connection by thus storing all the received signals for later processing. This will allow a usage in the fields for tests and not public endeavours. The space segment includes satellites but is not limited to this use. The required beacon signal can be transmitted from other vessels like high altitude balloons, flight drones or cars as well. The required beacon signal can be either transmitted by another hardware we offer or by just using the required parameters for transmission and identification signal to be used by the ground station system. So we provide both, the hardware and the standard. III.II DGSN Stations Ground stations devices are small modular systems to be either connected to a personal computer or as standalone devices. The device will include the basic unit that will use GNSS receivers to receive GNSS time IAC-13-B4,3,11,x17101 signals that will be either used to locate the stations position and to correlate the measured data with a high precision time source. The measuring infrastructure will allow modules to be plugged into the system and the first application will be the beacon signal receiver, but more measurements modules can be plugged in for various sensoric purposes. In addition to the beacon signal receiver module, the antenna periphery can be attached. The on-board data handling and network communication unit will handle the data traffic in the ground station and allows connecting an internal storage unit or a connection to an external computer. DGSN concept is divided in space, ground and network segments. IV. EXISTING GROUND STATION NETWORKS DGSN uses a new participatory approach and covers a niche that is not covered by existing ground stations networks. For tracking satellites NORAD (North American Aerospace Defense Command) offers Two-line element set (TLE) of satellites and space vehicles. NORAD uses active and passive tracking methods and regularly updates the public TLE data-base. The Doppler Orbitography and Radiopositioning Integrated by Satellite (DORIS) by the French Space Agency (CNES) uses beacon signals by satellites for positioning. A beacon signal is send out by ground stations and the frequency shift that occurs due to the high relative velocity is processed on-board the satellite. By the observation the satellite orbit, ground positions and other parameters can be derived. 50-60 stations are distributed over the Earth. These stations only emit signals but do not receive signals. On the one hand the installation of stations only requires electricity and the positioning can be processed by the satellite, on the other hand no payload data can be collected thus now position data is sent to the owner via the network. Furthermore it requires a reception of a bi-frequency 0.4 GHz and 2 GHz signal and additional processing power for the positioning. This is an additional challenge for small and smaller satellites. A combination of tracking and communication is provided by the Deep Space Network (DSN, NASA) and the European Space Tracking (ESTRACK, ESA) network. Both networks are set up by three ground stations with deep space antennas (35m diameter) providing a hemispherical signal propagation and reception. In addition to the main stations additional ground stations for communication with lower orbits (15m diameter antennas) are used. DSN is mainly used for missions beyond 16.000 km height. In this altitude the satellite is always in view of one of the three main DSN ground station antennas. For missions in lower orbits other NASA ground stations or the tracking and Page 3 of 11
  4. 4. 64th International Astronautical Congress, Beijing, China. This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License, 2013. data relay satellites (TDRS) are used. A similar approach is used by ESTRACK. DSN and ESTRACK are services are beyond budgets of small satellites. For university and amateur radio satellites the Global Educational Network for Satellite Operations (GENSO) [5] was started. GENSO is a worldwide network of ground stations that are connected via the internet. The stations can be actively controlled via a main server by an operator. During this phase each ground station will grant control access and will remotely communicate to the satellite if it is in view. Due to the active communication to the satellite the network is composed of only the amateur radio community because operational licensing is required. The satellite mission operator requests communication time via GENSO’s Authentication Server (AUS) and will get access to the network and can remote control each Ground Station Server (GSS) and the motorized antennas via the Mission Control Client (MCC). There are currently 25 ground stations inside the GENSO network. This is a time sharing approach by existing semi-professional ground stations offering an efficient usage of the stations. This method offers synergetic effects to each member of the community due to the fact that during the time the own satellite is out-of-view for communication another member is allowed to use the own station for their mission (Fig. 2). Fig. 2: Comparison between DGSN and GENSO network topology. No RX/TX for DGSN participants, only via mission control terminal. V. POSITIONING The main task of the DGSN is the tracking of beacon signals of satellites. This only requires low datarate and a regular emission of the signal for obtaining the orbital parameters. Trilateration is used which allows pseudo-ranging the signal’s origin position. V.I GNSS Relation Each ground station is equipped with GNSS modules. GNS-services like GPS and Galileo provides an independent and high accuracy time source that is globally available. This is used for the station’s system clocks. Furthermore GNSS can also automatically IAC-13-B4,3,11,x17101 provide the ground station location that with minimum manual set-up to make it user-friendly and to allow higher accuracy measurements. The ground station location is used to determine all station positions in the DGSN. The time source is used to be correlated with the satellite beacon signal when received. The same signal received at minimum five stations can be used to trilaterate the signals origin position. In this way the beacon and thus the satellite can be tracked and with a global network of many more ground station will allow it to permanently track it. All in all a global available and precise time source is essential and this can be only provided with GNSS. This gives the opportunity to use the ground station hardware as part of an internet connected sensor grid with thousands of sensor nodes, or even without as a 5 node sensor grid storing on internal memory units. V.II Trilateration The first phase of DGSN stations will use trilateration for determination of the signal’s original location. This is due to the fact to use a simple and proven method for positioning. Furthermore it allows a simpler design of the ground station and antennas. Additional features like frequency and phase shift measurement will be updated by software and due to the modular base design of the hardware. As an opensource project the application of this method is licensing-free and also allows modification by the community. DGSN implements a “reverse-GPS” determination of position. Instead of sending out signals from satellites with known orbital positions and calculating the unknown position of the GNSS receiver, a beacon signal with a unique time-stamp from an unknown position is send out and received at ground stations with known positions. The advantage of pseudoranging is applied so that accuracy errors in the measured reception time at each ground station can be determined. Due to offset of the satellite’s clock, atmospheric effects on the signal propagation path, errors in global time synchronization of the stations or the stations system delays, variations in the reception time occurs. So the satellite’s clock cannot be trusted and by using pseudoranging the beacon’s position and the time error can be computed accurately. This is based on the satellite’s time-stamp sent with the beacon signal and marking the send-out time and the reception time at the ground station. The signal path length is obtained by multiplying the speed of light by the time the signal has taken from the beacon to the receiving ground station. By using four stations the problem of accuracy errors will be transferred to spheres with the diameter of the computed paths. The location of the beacon signal is expected to be on each sphere’s surface and the three-dimensional coordinates can be Page 4 of 11
  5. 5. 64th International Astronautical Congress, Beijing, China. This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License, 2013. computed with four spheres under the premise of elimination of one of the two results. By using five stations the position can be determined explicitly. 100 km on Earth’s surface, variations in station system delays between 16 ms and 100 ms and the location (x,y,z) of the beacon origin in this area and up to altitudes of 1000 km were generated. With this the reception times at each ground stations is computed on basis of the propagation paths from the beacon location to each ground stations. The number of stations ranged from 5 to 15. Fig. 3: Trilateration in 2D plane with 3 circles (P1-3) resulting in beacon circle B with radius rB due to time error. A solution for this problem was firstly posed by Apollonius of Perga (262 BC – 190 BC) In case of errors in the time-accuracy the position of the beacon signal does not lie on the surface of all spheres but instead in the center of an additional sphere that is tangent to all other spheres. The radius of this sphere is the average time error of the position system. The 3D-positioning is preceded by solving a set of four linear equations for four ground stations or 5 linear equations for 5 stations. The last option eliminates the false solution by using four equations for four unknowns (determined system). [1] By using additional ground stations and thus spheres the system of linear equation is overdetermined. This is used for smoothing time-error effects on the calculation by either using a method for directly solving the overdetermined system or by reducing the calculation to a determined system and computing a set of ground stations by means of variations and combinatorial methods. V.III Positioning Software The “Lone Pseudoranger” software was developed and includes both methods. The behaviour of the algorithms was tested with a set of randomized input parameters generated by a Monte Carlo method. The locations (x,y,z) of ground stations within a 100 km x IAC-13-B4,3,11,x17101 Fig. 4: Lone Pseudoranger with mode 0 (top) and mode 1 (bottom) The behaviour of the determined system mode (mode 0) with four stations identifies that an additional determination methods needed to be included to distinguish between the two correct results of the quadratic function of the system of linear sphere equations. Because only four stations are used these can be combinatorial varied on the base of 15 stations. Even though each iteration will result in one result that is faroff the original beacon, the other one is assumed to be near to it. This condition can be used for filtering and cluster resulting positions that are in close proximity and filtering-out half of all results that are highly distributed. The behaviour of the overdetermined system mode (mode 1) was applied to all 15 stations. It only results in a single solution for the beacon position. The algorithm is faster but is very sensitive to variation in internal system delays due to the hardware. As an open-source project the user could integrate different microcontrollers and processors which results in different signal processing times and overall delay. With a delay range between typical values of 16 ms and 100 Page 5 of 11
  6. 6. 64th International Astronautical Congress, Beijing, China. This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License, 2013. ms the computed position are closer to the beacon origin when all station delays are the same. With a high distribution between minimum and maximum delays the computed position is further away. This is the result of solving all 15 equations with one overdetermined system of linear equation. If one or two stations inject different delay times in a system with equal delay times, the error is applied in an averaged way resulting in higher position differences. As a solution a overdetermined system is used to benefit by single solutions but with less stations than available. The remaining stations allow a combinatorial computation resulting in a set of positions. A filter detects the position with the furthest distance to the main cluster. V.IV Clustering Clustering the computed positions is one benefit of the distributed ground station network because there are more stations available then needed for trilateration. Lone Pseudoranger uses a simple sorting algorithm. The positions are computed by trilateration. If there are more than five stations and thus computed beacon position distances from each point to every other point is calculated. The next steps search inside the matrix of distances and assign the connection to the corresponding points from the shortest to the longest distance until each point is connected to at least one other position (Fig. 5). The next step includes a graph check whether all points are connected to each other or if there are clusters. If there are more than one cluster the cluster with most positions are determined. It is assumed that the cluster with the most position points is near to the real position and the final result is the geometric center of this cluster (Fig. 6). Fig. 6: Allocation of computed positions by clustering V.V Numerical Results The results of the randomized inputs showed four behaviours of the computed position (Fig. 7). 1) When the ground stations are located and only the beacon position is changed, then the higher the altitude of the beacon is, the smaller the field-of-view angle will be and the higher the inaccuracy dR. So for determination of satellites in different orbits, each orbits requires an adapted distribution density of ground stations. 2) When all ground stations system delay times are equal the highest accuracy is reached. This is independently of the value (16 ms to 100 ms) because the time-error is computed. In case each delay time is different the position accuracy decreases. This is an important fact for the open-source ground-station design. This can be respected by measured system delay times that are also included in time correction. 3) The number of ground stations receiving the same beacon signal shall be high. There was no obvious trend of converging to the minimum distance to the real value by increasing numbers of stations. The distribution of stations has a big influence on the accuracy and thus the number of stations. 4) Mode 1 showed to obtain better results than mode 0 with few ground station numbers (5 < stations < 10). With station numbers ≥10 the accuracy of mode 0 increases to the range of mode 1. So a combination of both modes with variable station numbers in the system of linear equations is applied as an additional position evaluation. Fig. 5: Basic steps of clustering algorithm IAC-13-B4,3,11,x17101 Page 6 of 11
  7. 7. 64th International Astronautical Congress, Beijing, China. This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License, 2013. VII.I Frequency Allocation The selected frequency ranges are amateur radio bands. This is required to open DGSN for everybody, including the amateur radio community. The candidate bands for DGSN are 70 cm, 23 cm and 13 cm bands (0.42, 1.24, 2.4 GHz). The 13 cm band is a popular band for small satellites and is also used by VERDE and FlyingLaptop. VI. GLOBAL DATA DUMP COMMUNICATION The global reception of beacon signals also offers the opportunity of receiving housekeeping payload data. With the simple antenna design of phase 1 only low data-rates of >2 kbit/s are set. This is adequate for essential data with high priority (Table 1). VERDE Data-Rate Contact Data High-Data 1200 kbit/s 10 min 90 Mbyte Beacon 2 kbit/s 100 min 1.5 Mbyte Table 1: VERDE sat communication links from 575 km SSO (100 minutes orbit). High-data rate in 13 cm band, Constellation in 23 cm band. The data-rate is due to omnidirectional antennas. For simplicity no motorized high-gain antennas are used. This is an option for later updates of DGSN. Furthermore the data-dump mode is a one way communication. The ground stations will only receive the data package and send it to the central server via the internet. The satellite owner accesses the data on the server and check for integrity. Missing or damaged packages will be ordered to be resend via the satellite owner’s main communication terminals. The satellite will then resend these packages as ordered via the DGSN. VII. GROUND STATION CONSIDERATIONS Regulations The main regulation will be amateur radio licensing. DGSN is a participatory project and it is open for everybody. This requires that the ground stations do not emit signals. So the ground stations in phase 0 will only include “listening” mode and can only receive signals and will forward it to the central server. For a later phase and for authorized users, ground stations could also be used for two way communications. IAC-13-B4,3,11,x17101 [2] The propagation path S is adapted according to the elevation angle and longer distances to the satellite in orbit. The main influence is the system noise. It is higher for the 70 cm band resulting in bigger system margins of higher frequency in 13 cm band (221 K @ 0.4 GHz, 135 K @ 2-12 GHz). By using the elevation angle the solid angle of the propagation cone can be determined. The maximum cone angle is reached when the system margin drops below 4 dB (Fig. 8). 2.4GHz@ 16,0 0.5kbit/s 1.2GHz@ 12,0 0.5kbit/s 0.4GHz@ 0.5kbit/s 8,0 2.4GHz@ 1kbit/s 1.2GHz@ 4,0 1kbit/s 0.4GHz@ 0,0 1kbit/s 0 30 60 90 2.4GHz@ 2kbit/s -4,0 1.2GHz@ 2kbit/s -8,0 0.4GHz@ elevation angle [°] 2kbit/s system margin [dB] Fig. 7: Trilateration accuracy over ground station number. The accuracy is presented as the difference between computed and real beacon position. VII.II Station Density The radio link budgets for up and downlinks are modelled after “Space Mission Analysis and Design” by W. J. Larson and J. R. Wertz [6] [7]. Fig. 8: System margin over elevation angle. Graph shows the system margin for links with 70, 23, 13 cm bands with 0.5, 1.0 and 2.0 kbit/s in correlation to the elevation. The signal path length to VERDE in 575 km varies according to elevation angle. The angle defines the chain of ground stations that is required for a full coverage of 360° on the equator (longitude). By expanding to the latitude direction the overall number of stations is obtained. However the stations will be deployed on land the density of stations increases. The number of ground stations of 365 for a beacon link with 2 kbit/s in 2.4 GHz is expected to be realistic Page 7 of 11
  8. 8. 64th International Astronautical Congress, Beijing, China. This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License, 2013. (Fig. 9). For later phases the data-rate shall be increased. And even though the data-rate is low compared to other shared infrastructure concepts, it allows the simultaneous usage of different satellite operators. In those concepts the contact area is wider than those of the DGSN stations. This also means that if an operated allocates a ground station server this station is blocked, no other operator can use it, and there is a high chance of the satellite flying by in exactly this grid section. With DGSN there are more nodes and it increases the number of simultaneously received satellites. In this way the effective data-rate of all traffic also increases. 1000 800 600 400 200 0 1013 network and only process the demanded data of a demanded satellite. The processing and tracking is possible with the id but it requires additional steps. The beacon signal frame also allows the satellite owner to use the optional part of the frame for data that is not required for tracking. This is the case for the datadump mode. The package size should allow to be broadcasted and receivable with low system margin and should respect the current weather situation. This data could be further data for the use of tracking correction like IMU position data for correlation with the computed values or frequency and phase information of the transmitter. And finally the payload data for the global dump mode is included. The satellite operator is completely free to transmit data in this optional frame segment. 613 72 98 128 145 221 288 365 Fig. 10: Beacon signal example frame Fig. 9: Number of deployed station for global coverage. VIII. BEACON AND GROUND STATION ARCHITECTURE VIII.I Beacon For the first phase of DGSN the beacon signal is a simple digital stream and alternatives like pseudorandom codes and more advanced concepts for real-time reception are proposed for later updates. The signal itself is a periodically transmitted frame and is splitted in a mandatory section containing the essential starting marker, unique event and satellite id, and the optional part where additional payload data and tracking correction data can be transmitted (Fig. 10). The mandatory part starts with a marker for addressing the following frame and it can be used for synchronization. Additionally the marker can be used for time correlation at the ground station. After that a unique event is allocated. This unique event is required for the detection of the same event at different locations received at different times. If the repetition of frame without a unique event is to fast the received signal event is not distinguishable and accuracy of detection is decreased or not possible. The unique event can be satellite data, a random number, or a time-code. The last option allows detection of clock offsets and errors. And the first option increases usage of the bandwidth. The satellite id is required for the following signal pre- and processing to filter all received data in the IAC-13-B4,3,11,x17101 VIII.II Ground Station The DGSN concept is based on time correlation provided by a GNSS source (currently GPS, later Galileo and Glonass) with an event signal (Fig. 11). This event can be a satellite beacon or any other beacon with the beacon signal protocol or others (ADS-B, etc.). For a flexible use of time source and even event signals the ground station concept is modular allowing the user to expand and use the DGSN modules as a basis and any other compatible module. The received signal is processed in the receiver modules and transferred to the bus connector. The event data is than correlated with the received time by the time source board both are collected by the data collector and finally transferred to the ground station server of the participants operating the node. Then the data is submitted via the internet to the central Constellation server. Fig. 11: Allocation of computed positions by clustering Page 8 of 11
  9. 9. 64th International Astronautical Congress, Beijing, China. This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License, 2013. The basic DGSN ground station is equipped with a satellite event receiver (SER) and the time board (TB) (Fig, 12). At first the time board is activated via the data bus triggering a microcontroller to supervise and configure the GPS module. The GPS module delivers the location of the ground station and then provides the one pulse per second (1PPS) time signal to the PhaseLocked Loop / Direct Digital Synthesizer (PLL/DDS). The PPL/DDS synchronizes to the 1PPS and multiplies the frequency with a stable and higher frequency (e.g. 10 MHz)- Both signals are transferred to the latched counter that calculates the current system time based on the 1PPS GPS time and increases the accuracy by using PLL time for counting the intermediate time steps between two 1PPS signals. The system time is then constantly buffered in the latched counter block. Fig. 13: Proposed test configuration of five mobile ground stations for ESRANGE. The open-source approach allows application outside DGSN and is discussed to be used for FREDE experiment on BEXUS high altitude balloon flight. Fig. 12: Example of setup with a high precision time source and measurement for ground station The satellite beacon signal is received and injected into the bus-connector. The event signal is detected and the latched signal triggers the read-out of the latched counter. The currently buffered latched time count value is injected into the bus-connector and both signals are correlated and transferred to the data-collector. For this the time board hast to permanently hold available the updated latched counter to be prepared for instantaneously providing the event time when the event signal is detected. VIII.III Configuration The beacon and ground station configuration is designed to function with and without network connection. The principle of five stations for tracking beacon signals can be used in the DGSN project but also on the fields for mobile and remote experiments like on ESA ESRANGE, Sweden (Fig. 13). The “fast deployable antenna” concept is applied where the data is stored on ground station on-board memory that needs to be collected and analysed after the experiments activities. IAC-13-B4,3,11,x17101 IX. NETWORK SEGMENT INFRASTRUCTURE The open-source approach allows free reproducing and utilization of the hardware and the user can also setup their own network. For DGSN the Constellation platform is used for the open tracking and communication service. IX.I Constellation Distributed Computing Platform Constellation is a distributed computing supercomputer. More than 7000 volunteers share their idle PC time at home and for a virtual super computer via the internet (Table 2). Constellation has been active since 2011 and their users solve numerical aerospace problems. On each of their computer the BOINC (Berkeley Open Infrastructure for Network Computing) is installed. It can be attached to various citizen science projects and computing their workunits. Constellation’s user basis is spread worldwide and shall serve as nodes for ground stations. The average user shares their computing capacity for five hours per day as BOINC runs in the background. It is expected to motivate 5 % of the user base to participate in DGSN and buying a ground station kit or assembling components. The ground station is designed to be installed easily and the omnidirectional antenna allows deployment in various environments (windows, balconies, front gardens). Page 9 of 11
  10. 10. 64th International Astronautical Congress, Beijing, China. This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License, 2013. Users Countries TeraFLOPS Constellation 7.848 108 3.957 Table 2: Constellation Distributed Computing Grid. [8] Right away Constellation offers computing capacity for signal processing the received data as well as determination of the two-line element set. The satellite operator can receive final track products. Furthermore the operator is performing outreach by actively involving the public in the satellite mission. It can be used to inform about the mission, to benefit from developments (software and hardware) by the community, providing synergies between the operator and community and also for recruiting outstanding, skilled users. IX.II Active Distributed Sensoring Projects Besides distributed computing there is also distributed sensoring. For this the user only provides a minimum computing capacity and shares the idle time for sensoring with small devices plugged to the computer. During this time sensor data is recorded and also sent to the central server. In this way regional and global maps are generated (Table 3). RadioActive@home and Quake Catcher Network are BOINC projects and using the client for sensor operations and data-management. The first project acquires data of environmental background radiation and the second measures Earth quakes with three-axes accelerometers. The Air Quality Egg and Blitzortung community use special sensors and own software for sensoring and data transfer. The prize of their hardware is in range of the expected hardware cost per unit of DGSN. So a community size of 400 is plausible. Project Nodes Price/Unit RadioActive@home[3] 386 37 € Air Quality Egg [9] 650 100 € [10] 749 200 € Quake Catcher Network[3] 4397 35 € DGSN >400 200 € (est.) Table 3: Citizen science projects with distributed sensors. The number of nodes and price per sensor unit is presented (28.Aug.2013) in correlation to the expected DGSN nodes and estimated price per unit. The main drivers of people participating in these projects are a general interest in the specific field of science and out of conviction to make a small but important contribution for a project where everyone benefits in everyday life. This can be transferred to space applications and research offering an opportunity for DGSN. IAC-13-B4,3,11,x17101 X. PERSPECTIVE X.I Test Campaign A first set of 5 prototypes will be built to be able to track with trilateration ("reverse-GPS"). These stations will be equipped with modules for ADS-B to test the trilateration software with aircraft data system signals to use their transmitted position in ADS-B for validation. After those tests the beacon receiver module will be tested to be able to receive our own beacon signals on ground "fox hunt" tests and on high altitude balloons. This will allow a stepwise approach to reception of satellite beacon signals. The last test phase will include a deployment of 50 post-prototype ground station devices in across Europe to test the global coverage of beacon signals. • ground (fox hunt) • air (quadrocopter, ADS-B) • high altitude (REXUS/BEXUS [11]) • space (small satellites, ArduSat, FLP) X.II Target Group The target group spreads among small, nano and cube satellites owner. They are mainly operated by universities and the radio amateur community. Those groups have a need for global and cheap communication links to their satellites and payloads and also rely on positioning of their satellites. The access to the DGSN is meant to be open for everyone to have a wide range of possible users of the network. In this sector there is a market for services to integrate the needed hardware and software for their satellites. Another market sector is the ground station hardware itself, which has to be built and deployed. This relies on citizen scientists and a huge basis of small and cheap devices. The biggest market will be derivates from sensor data and their derivates spawning into other sections besides the telemetry and tracking of satellites. A global sensor grid of ground stations could be used to measure GNSS accuracy and serve as a basis to improve the GNSS service itself (WAAS, EGNOS) [12]. The ground station can be used for augmented-GPS on ground when extended with transmission capacity. Tracking of beacon signals can serve tracking from weather balloons to skiers with functional clothing with build in beacons in avalanche situations. Or just for mapping in OpenStreetMap [13]. The generated data-basis has the potential to spawn this extended uses and the hardware can be extended and upgraded for more measurement purposes. The big satellite owners already use their own infrastructures or pay a service like DORIS by CNES or use other sources by NORAD and ESTRACK. The DGSN concept can be an additional service for small players and can set up the first privately operated Page 10 of 11
  11. 11. 64th International Astronautical Congress, Beijing, China. This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License, 2013. service provider for tracking and for small telemetry reception. X.III Open Access The business case includes an open access strategy and this includes a citizen science participation, a standard setting and providing of hardware with later derivates of services. The DSGN must be open accessible for a broad acceptance of clients and their needs. There is a need and thus a market for global tracking and data reception service for small satellites and other mobile targets. The aim is to provide them with a service they need in combination with participants, which have not been part in space based operations before and thus creating a community and thus a new small market. The satellite owner gets what he wants, the citizen scientists can help by voluntarily work and be credited and the hard- and software segment can be pushed that increases sales. And the additionally spawned services like an additional, but private, satellite tracking services among NORAD, DORIS & Co. and new services like automated mapping for OpenStreetMap and creating data sets for further research gives possibilities. Potential customers are small satellite owners and services, like universities or even the German Aerospace Center DLR, where several small satellite projects are on their way and concepts for ground station networks exists (but not to such an extent and as an closed environment solution). REFERENCES [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] Fig. 14: small satellites interested in DGSN To name a few small satellite projects with their own solution for ground station networks • nano and small satellites by universities (RAX, AISat, FlyingLaptop) • satellite constellations and swarms (QB50 & GENSO) • re-entry vessels (MIRKA 2) • high altitude experiments (REXUS/BEXUS) • balloons (weather balloons, BEXUS, FREDE experiment) • planes and drones (ADS-B, Stuttgarter Adler) • sensoring platform (thunderstorm-, flash-, nuclear detonation detection) • GNSS quality measuring (WAAS, EGNOS) • safety of life (avalanche) IAC-13-B4,3,11,x17101 [11] [12] [13] IRS, University of Stuttgart, "Small Satellite Project," [Online]. Available: nanosatisfy, „ArduSat,“ [Online]. Available: "Project FREDE - Freon Decay Experiment," [Online]. Available: "Constellation Distributed Computing Platform," [Online]. Available: L. Mehnen, B. Preindl und S. Krinninger, „The Potential of Ground Station Networks like GENSO for Multi-Satellite Projects like QB50,“ in IAC 2010, Prague, CZ, 2010. Wertz, J. R.; Larson, W. J., Space Mission Analysis and Design, 3rd edition (ISBN:1881883108), Mcgraw-Hill, 1999. C. Nöldeke, Satellite Communications (ISBN: 3869914017), Monsenstein und Vannerdat, 2013. BOINCstats, "Project stats info," [Online]. Available: „AirQualityEgg - Environment Sensor,“ [Online]. Available:, "Participants," [Online]. Available: p?lang=en&page=4&subpage_0=10. REXUS/BEXUS, "Rocket and Balloon Experiments for University Students," [Online]. Available: T. Feuerle, GPSund Ionosphärenmonitoring mit Low-Cost GPSEmpfängern (ISBN:9783928628549), TU Braunschweig Campus Forschungsflughafen. OpenStreetMap, „The Free Wiki World Map,“ [Online]. Available: Page 11 of 11