MOSAIC: Mars Orbiting Satellites for Advanced Interplanetary Communication

Transcription

1 University of Illinois at Urbana-Champaign MOSAIC: Mars Orbiting Satellites for Advanced Interplanetary Communication Illinois Space Society Team Lead: Christopher Lorenz Team Mentor: Denis Curtin, Society of Satellite Professionals Alexandra Bacula Alexander Case Benjamin Collins Zachary Fester Brian Hardy Yukti Kathuria Andrew Koehler Steven Macenski Joseph Miceli Jordan Murphy Jeffrey Pekosh Kaushik Ponnapalli Lui Suzuki Kelsey White

2 Table of Contents Table of Contents... 1 List of Tables ) Introduction ) Proposed Architecture ) Phase 1 Overview ) Phase 2 Overview ) Phase 3 Overview ) Ground Infrastructure and Operations ) Phase 1 Small Satellite Constellation ) Small Satellite Deployment ) Constellation Configuration ) Launch and Orbital Insertion ) Satellite Hardware ) Communications Payloads ) Satellite Bus Selection ) Phase 2 High Altitude SmallSats and Laser Relay Satellite ) Orbital Mechanics ) SmallSats ) Relay Satellites ) Satellite Hardware ) Communications Payloads ) Satellite Bus Selection ) Production Schedule and Launch Frequency ) SmallSat Launch Analysis ) LACOR Launch Frequency ) Phase 3 Solar Eclipse Relay Satellite ) Orbital Mechanics ) Satellite Hardware ) Communications Hardware Illinois Space Society 1

3 5.2.2) Satellite Bus ) Ground Infrastructure and Operations ) Site Locations for Optical Ground Network ) Technical Specifications for Ground Telescopes ) Programmatic Considerations ) Architecture Cost Analysis ) Risk Analysis ) Conclusions Appendix A: References Appendix B: Nomenclature Appendix C: Uplink Budgets Appendix D: Laser Communication Additional Values Appendix E: Launch Schedule for All Satellites Illinois Space Society 2

4 List of Figures Figure 1: Visual overview of communication for Phase Figure 2: Visual overview of communication for Phase Figure 3: Visual overview of communication during eclipse for Phase 3 (non-eclipse periods will be identical to Phase 2) Figure 4. Plot showing the percentage of Mars surface covered by a single communication satellite versus altitude Figure 5. EMTG plot of the 2024 Mars transfer orbit Figure 6. Pareto surface of the 2037 Earth-Mars transfer trajectory, showing departure and arrival C3 values versus the required correction ΔV on the vertical axis Figure 7. Plot of ΔV used for orbit phasing versus phasing time required for Phase Figure 8: An illustration of the final position of each satellite in a plane after orbit phasing Figure 9: A TerraSense 300 seen in orbit around Mars with its communications payloads Figure 10: TerraSense 300 satellites (black) arranged around the ESPA Grande ring (orange) with a STAR 63F (green) attached shown inside an Atlas V 5 m fairing Figure 11. ΔV versus time required for phasing maneuver for Phase Figure 12: Model of LACOR satellite with laser communication terminal shown on the top of the bus and the X-band antenna on the bottom Figure 13: LACOR shown in the launch configuration with the high gain antenna located at the top of the diagram Figure 14: Plot showing a successful example run of the lifetime analysis. The dotted red line shows the required number of satellites to achieve 100% coverage Figure 15: Mapped site locations for an LDOS with eight ground stations (world map background from [63]) Figure 16: Cassegrain configuration for 10 m optical receiving telescopes (diagram from [64]) 35 Figure 17: Optical signal detection system for 10 m receiving telescopes (diagram from [64]).. 35 Figure 18: MOSAIC annual cost for the timeframe Illinois Space Society 3

5 List of Tables Table 1. Table of Launch Vehicle and Insertion Stage Required for All Launch Windows for the Phase 1 Altitude Table 2: Downlink Budget for Robotic Rovers to Low Altitude SmallSats Table 3: SmallSat Communications Payload Mass Budget Table 4: Downlink Budget for SmallSats Direct to Earth via a 34 m DSN Antenna Table 5: SmallSat Hardware Requirements Table 6. Trade Study on the Small Satellite Bus Structures Table 7. Table of Launch Vehicle and Insertion Stage Required for Various Launch Windows for the Phase 2 Altitude Table 8: Downlink Budget for a Manned Base to the SmallSats Table 9: Downlink Budget for SmallSats to the Sun-Mars L1 Relay Satellite Table 10: Laser Communication Data Rate Calculation Values from the L1 Relay to Earth Table 11: LACOR Communication Payload Mass Budget Table 12: LACOR Hardware Requirements Table 13: Relay Satellite Bus Trade Study Table 14: SER Communications Payload Mass Budget Table 15: SER Hardware Requirements Table 16: Detailed Site Locations for an Eight-Station LDOS (reproduced from [62]) Table 17: Definitions of Probabilities and Impact Levels Table 18: Definitions of Architecture Risk Colorations Table 19: Launch and Deployment Risks to the MOSAIC Architecture Table 20: Communication System Risks for the MOSAIC Architecture Table 21: Uplink Budget for DSN to SmallSat Table 22: Uplink Budget for SmallSat to Rover Table 23: Uplink Budget for Sun-Mars L1 Relay to SmallSats Table 24: Uplink Budget for SmallSat to Manned Base Table 25: Downlink for LACOR to SER laser communication Table 26: Uplink for SER to LACOR Laser Communication Table 27: Downlink from SER to the Earth Ground Stations Illinois Space Society 4

6 1) Introduction NASA s recent focus for human spaceflight has been a build-up of capability towards landing humans on Mars. The advancement of key technologies currently underway has a final goal of being able to provide a sustained human presence on the Red Planet. This final goal will need to be supported by a complete infrastructure, both in space and on the ground. The telecommunications capability required for a manned Mars mission will far exceed that of any robotic endeavor to date. This report outlines a three phase architecture with corresponding ground infrastructure, in which a slow build-up of assets will lead to an effective network for Mars communications. This system will support a continued robotic presence and the eventual arrival of humans with high data rate communications capability and multiple layers of redundancy. 2) Proposed Architecture 2.1) Phase 1 Overview Phase 1 of this architecture will allow for up to 61.2% coverage of Mars by 2025 at data rates averaging about 550 kbps per satellite. This segment of the architecture will see sets of four small satellites (SmallSats) launched to Mars about every two years into low orbits (1507 km altitude). These satellites will communicate direct to Earth using X-band communications as shown in Figure 1. The relatively low data rate capability during this period will be sufficient to support continued robotic exploration. The TerraSense 300 commercial SmallSat bus has been selected to permit the mass production of these satellites, allowing a continuous constellation of at least six to be active far into the future. Figure 1: Visual overview of communication for Phase 1. Illinois Space Society 5

7 2.2) Phase 2 Overview Phase 2 of the architecture will see SmallSats identical to those in Phase 1 launched into higher orbits (10,716 km altitude) to provide 100% continuous coverage of the Martian surface in support of manned exploration. These satellites will begin launching in 2029 and will provide a link for a sustained robotic presence. They will launch four at a time every synodic period (780 days), allowing a consistent build-up and maintenance of capability. However, these satellites will rely on the rovers and landers to carry a stronger communications link due to their high altitude. This trade-off allows for full coverage of the planet at high data rates for potential human visitors, at the cost of stronger communication constraints on robotic missions. Figure 2: Visual overview of communication for Phase 2. In addition to these high altitude SmallSats, a pair of laser communication relay satellites (LACOR) will be sent to the Sun-Mars L1 point to provide a high data rate upgrade to the system beginning in Communications links during this period will rely on X-band from the SmallSats to LACOR and then optical communication from the relay to Earth as shown in Figure 2. These larger relay satellites will enable data rates of up to 80 Mbps from two locations on the Martian surface continuously. These satellites will be constructed using a commercial off the shelf (COTS) bus, the GEOStar-2 built by Orbital ATK. 2.3) Phase 3 Overview Phase 3 of the MOSAIC architecture is designed to fill the coverage gap that Phase 2 encounters when the Sun blocks communication between Earth and Mars. The final addition of a Solar Eclipse Relay (SER) satellite at the Sun-Earth L4 point will provide continuous coverage of 100% of the Martian surface while the Sun is directly between Earth and Mars, albeit at a lower data rate. Instead of communicating directly to Earth, LACOR will communicate to SER, which in turn relays data back to Earth. This added link to the communications chain seen in Figure 3 lasts about 21.5 days and occurs every 780 days. This final addition (launched in 2033) completes Illinois Space Society 6

8 the architecture, providing a reliable and efficient link between Earth and Mars to support a sustained human presence on the surface. Figure 3: Visual overview of communication during eclipse for Phase 3 (non-eclipse periods will be identical to Phase 2). Continued support will be given to the Mars orbiting constellation and Sun-Mars L1 relay satellites in the form of continual replacements and spares. Upgrades to this network will be made as required to provide additional data return capability, utilizing any new technologies that may exist later into this exploration phase. The mission architecture has been developed out to 2045, but the opportunity to extend it for a permanent human presence has also been outlined. 2.4) Ground Infrastructure and Operations Proper ground infrastructure is vital for each phase of the MOSAIC mission. In Phase 1, the pre-existing Deep Space Network (DSN) will support X-band transmissions between Earth and the network of SmallSats around Mars. The DSN is an international ground-based communications system, providing coverage via its three ground stations spread around the globe. With antennas ranging in size from 34 meters to 70 meters, the network enables NASA and other space agencies to communicate with probes throughout the Solar System. The DSN s link availability is also excellent, with system availability consistently above 95% and often above 99%. [1] Phase 1 of MOSAIC will utilize the DSN s 34 m antennas as both receiver and transmit terminals for communicating with the SmallSats. To support optical communications in Phases 2 and 3, a new optical ground network will need to be constructed. This network will consist of eight ground stations spread evenly around the world, with each site specifically chosen to maximize link availability. Each station will have an uplink and downlink terminal designed for optical communications. More detail on the planned optical ground network can be found in Section 6. Illinois Space Society 7

9 3) Phase 1 Small Satellite Constellation Small satellites are seeing increased use in space exploration, as spacecraft components are being miniaturized. In the case of the MOSAIC architecture, SmallSats enable a lower cost option for data relay than larger satellite busses. By including many SmallSats, cost as well as risk decrease due to spacecraft redundancy. This lower risk profile is critical when considering the implications of human missions to Mars. All manned Mars architectures feature costs in the tens or hundreds of billions of dollars, so maximizing the scientific potential of the missions via reliable communication is critical. 3.1) Small Satellite Deployment 3.1.1) Constellation Configuration The initial SmallSats launched to Mars will be launched into a 55 degree inclination, 1,507 km altitude circular orbit. They will be launched in batches of four, into planes in which each satellite will be separated by 90 degrees true anomaly in their final configuration. This 55 degree inclination was chosen for the Phase 1 satellites as it will provide better coverage near equatorial latitudes than near polar orbits, while also providing some coverage at higher latitudes. The percentage of the surface covered by a single communication satellite is shown below in Figure 4. Figure 4. Plot showing the percentage of Mars surface covered by a single communication satellite versus altitude. At an altitude of 1,507 km, each SmallSat will provide coverage over 15.3% of the Martian surface, with each satellite situated at this altitude so that it will be able to communicate with two of the three other satellites. This ensures that any object that can communicate with any of the SmallSats can have its signal relayed back to Earth, even if Mars eclipses the SmallSat that is Illinois Space Society 8

10 communicating to the ground. Combining the 15.3% coverage per satellite with the inter-satellite communication gives a figure of 61.2% coverage, as the coverage patterns do not overlap. This figure will ensure that the 25% coverage goal can be more than met with two SmallSats during the Phase 1 portion of the mission, along with two additional SmallSats for redundancy ) Launch and Orbital Insertion The SmallSat transfers to Mars were calculated using the trajectory optimization software Evolutionary Mission Trajectory Generator (EMTG), using an MGA-DSM impulsive approximation. This software is an open source tool currently under development at NASA s Goddard Space Flight Center. [2] The tool is a medium fidelity trajectory optimizer that relies on random sampling of many thousands of feasible trajectories to find optimal solutions. The software is used by the Mission Design Branch of NASA Goddard to develop the agency s revolutionary mission concepts. Figure 5. EMTG plot of the 2024 Mars transfer orbit. Using this software, an Earth-Mars transfer was created such that any correction burns required would be minimized. The result is a trajectory that requires a minimal amount of ΔV, similar to the idealized Hohmann transfer. This resemblance can be seen above in Figure 5. This will allow the transfer vehicle to consist of a relatively dumb set of components. In addition, EMTG s genetic algorithm was used to vary the arrival and departure C3 values in order to minimize the insertion ΔV and to minimize the size of the launch vehicle and propulsive stage required. An example of the resulting Pareto surface is shown below in Figure 6. The ideal Mars transfer trajectory depends on which launch window to Mars is being considered. Each launch window was considered over a range of departure and arrival C3 values that were achievable with Atlas V variants and ATK STAR solid rocket motors that would be able to carry 4 SmallSats into the required Martian communications orbit. Illinois Space Society 9

11 Figure 6. Pareto surface of the 2037 Earth-Mars transfer trajectory, showing departure and arrival C3 values versus the required correction ΔV on the vertical axis. Each batch of SmallSats will be launched with some variant of the Atlas V, with the variant depending on the Earth-departure C3 required and the size of the insertion stage required to enter Martian orbit. The Atlas V was chosen for the launch vehicle as it has a payload capacity to Mars that is sufficient to carry the four SmallSats to Mars along with an insertion stage at a reasonable cost. An advantage of using an Atlas V is that for less ideal launch windows, a more powerful Atlas V variant can be used, while still allowing for the same ease of integration to the launch vehicle. This analysis assumed that the Atlas V will be used for all future launches of SmallSats, with the expectation that if the Vulcan rocket replaces the Atlas V in the 2020s, it would have comparable performance for an equal or lesser cost to the existing systems produced by ULA. [3] The Vulcan concept also currently has the same modular booster configurations as the Atlas V. The size of the insertion stage required to enter into Martian orbit depends on the altitude of the orbit chosen, and the Mars-arrival C3. Different sizes of Orbital ATK STAR motors will be utilized to conduct the Mars insertion maneuver. These motors were chosen because they are available in a number of different sizes, allowing for different launch trajectories which vary in ΔV requirements, as well as the fact that STAR motors are well proven and reliable. A table showing the different combinations of STAR motors and Atlas V variants required to meet different trajectory C3 requirements is shown below. Table 1 below contains the first three launches windows, which assume a payload mass of 1,307 kg, and a final orbit altitude of 1,507 km. Each insertion stage corresponds to a range of Mars arrival C3 values, and each launch vehicle utilized for a given pair of C3 values is denoted by color. The table optimizes the launch vehicle Illinois Space Society 10

12 used for a particular insertion stage, such that it uses the smallest launch vehicle that is capable of achieving that particular C3. Table 1. Table of Launch Vehicle and Insertion Stage Required for All Launch Windows for the Phase 1 Altitude C3 (km 2 /s 2 ) Earth-Departure Insertion Stage Mars-Arrival STAR 48B STAR 48A STAR 63F Atlas V Used After performing insertion, each satellite will separate, and perform any correction burns necessary to finalize its orbit and to obtain the correct phasing between each of the satellites. This phasing process will take 4.45 days to accomplish, and utilize 20 m/s of ΔV, out of the SmallSat budget of 90 m/s. This ΔV allotted to the orbit phasing was determined by plotting the time required for the spacecraft to reach the desired phasing, versus ΔV required. This plot, shown below in Figure 7, illustrates the diminishing returns that are experienced with high ΔV maneuvers, and the rapidly increasing transfer times associated with low ΔV ones. Illinois Space Society 11

13 Figure 7. Plot of ΔV used for orbit phasing versus phasing time required for Phase 1. For this reason, a maneuver requiring a reasonable 20 m/s was chosen, which would result in two satellites reaching their final positions in 4.45 days, and the other two in 1.48 days. The two faster satellites would each change their phase by 45 degrees, while the two that take longer require a phase change of 135 degrees. This is illustrated below in Figure 8, which shows the satellites in addition to the location of the booster before and after the phasing maneuvers. Figure 8: An illustration of the final position of each satellite in a plane after orbit phasing. Illinois Space Society 12

14 3.2) Satellite Hardware 3.2.1) Communications Payloads Communications calculations for the first phase of the mission assumed a rover on the surface of Mars communicating via a UHF link to the SmallSats, which act as a relay. The rover for this analysis is assumed to use the ElectraLite transceiver. This method of communication and transceiver hardware have been used for the Mars Science Laboratory. [4] Therefore, the small satellites are required to house a UHF transceiver and antenna for data relay. A pair of omnidirectional UHF antennae have been chosen, similar to those found on the Mars Odyssey orbiter. These antennae are an evolved version of those included on the Odyssey mission, and have a gain of 6.77 dbic. [5] Two are included to allow for full coverage regardless of spacecraft orientation. Although the antennae are omnidirectional, with only a single antenna the satellite bus would block transmissions in some orientations. The small satellites will use the Electra transceiver flown on the Mars Reconnaissance Orbiter. [6] This transceiver is capable of duplex UHF transmission and has flight heritage. A link budget analysis of this phase of the mission can be found in Table 2. The required value of Eb/No has been set to 0.31 for all link budget analyses in this paper, consistent with a bit error rate of 5e-3. This metric was used for the Cassini mission s data return and should be sufficient for Mars downlink. [7] The team set a standard of a 6 db link margin to ensure link viability as this concept will move from design into implementation. The link margin of 6.04 db exceeds this standard, yielding a viable link. An uplink analysis for this case can be found in Appendix C. Table 2: Downlink Budget for Robotic Rovers to Low Altitude SmallSats Specification Value Transmitter RF Power Output 13 Watts [8] Frequency MHz [8] Rover Antenna Gain 0.00 dbic [8] Miscellaneous Rover Losses db [8] Polarization Loss db Range to Spacecraft 1,507 km Path Loss db SmallSat Pointing Error 0.1 degrees SmallSat Pointing Loss db [6] SmallSat Antenna Receiving Gain 6.77 dbic [5] SmallSat System Noise Temperature 526 K [9] Demodulator Implementation Loss db [8] Data Rate 1,100 kbps Attained Eb/No 6.35 db Required Eb/No 0.31 db [7] Margin 6.04 db The analysis was performed assuming an altitude of 1,507 km above the Martian surface. The maximum data rate was found to be 1,100 kbps. Illinois Space Society 13

15 During Phase 1 of the MOSAIC architecture, the data received by these SmallSats must be directly transmitted back to Earth. X-band, Ka band, and laser communication options were considered for this system, but ultimately an X-band system was chosen. The system must be ready for launch by 2023, so a system with limited development is favorable. Laser communication was ruled out due to the difficult pointing constraints it would impose on the small spacecraft bus, which has limited pointing knowledge and accuracy. Laser communication at the scale required also has a relatively low Technology Readiness Level (TRL), making it difficult to implement for a near term mission. The X-band system will be able to draw heavily from hardware used in past missions, decreasing overall risk to the system over Ka-band options. Ka-band communication is also more strongly affected by Earth s atmosphere. [10] A 1.5 m parabolic antenna with 40.6 db of gain (similar to that of the NEAR spacecraft) will be included in the payload package to achieve the required downlink performance. [11] This high gain antenna will be rigidly attached to the spacecraft structure on the largest of the payload attachment faces. This rigidly attached antenna, along with the omnidirectional UHF antennae allow for greatly simplified spacecraft pointing operations. The spacecraft is simply required to point in the direction of Earth (or the relay for Phase 2) with no other pointing constraints. A COTS X-band transceiver has been chosen. The General Dynamics Small Deep-Space Transponder (SDST) is an off the shelf unit developed for communication with the Deep Space Network. [12] This transponder is similar to the unit flown on the Mars Reconnaissance Orbiter (MRO). [6] Also included in the communications system is a pair of travelling wave tube amplifiers (TWTAs). These units are capable of producing 100 W of RF transmitting power for an input of 172 W, and are based on the design of MRO. [6] Two are included for full redundancy, as these units are a common point of failure for spacecraft, including on the Voyager 1 spacecraft. [13] The UHF transponder has a receiving power requirement of 5 W, leading to a total payload power draw of 177 Watts. [14] Table 3: SmallSat Communications Payload Mass Budget Component Total Mass [kg] High Gain Antenna [11] 6.3 X-band Transponder [6] 3.2 X-band Power Converters 3.0 Travelling Wave Tube Amplifiers (2) 1.9 TWTA Switches and Other Hardware 3.1 Electra Radio 5.0 Low Gain Antenna (2) 2.2 Misc. Wiring 2.0 Subtotal 26.7 Contingency (43%) 11.5 Total 38.2 The total mass was found to be 38.2 kg including a 43% contingency as seen in Table 3. This level of contingency is the industry standard for pre-phase A conceptual designs. The maximum required power by the system is the input power into the SDST, 172 Watts. The link analysis performed for this segment of the communications architecture can be found in Table 4. The analysis assumes the satellite is communicating with a 34 m Deep Space Network Illinois Space Society 14

16 antenna while the Earth and Mars are at their maximum distance (about 2.68 AU). Other antenna diameters are available, but typical deep space missions use either a 34 m or 70 m dish for downlink purposes. Therefore, the 34 m dish and 2.68 AU distance give a worst case estimate for communications capability. The maximum downlink rate during this phase is 195 kbps for the worst case alignment of Earth and Mars. This scenario has a link margin of 6.03 db, proving link viability. When the same analysis was used to simulate the Mars Reconnaissance Orbiter, it yielded a data rate close to the published minimum value of 500 kbps. [6] Table 4: Downlink Budget for SmallSats Direct to Earth via a 34 m DSN Antenna Specification Value Transmitter RF Power Output 100 Watts [6] Frequency 8,400 MHz SmallSat Antenna Gain dbic Miscellaneous Losses db Polarization Loss db Range to Spacecraft 2.68 AU Path Loss db SmallSat Pointing Error 0.1 degrees SmallSat Pointing Loss db [6] DSN Antenna Receiving Gain dbic [15] DSN System Noise Temperature K Demodulator Implementation Loss db Data Rate 175 kbps Attained Eb/No 6.34 db Required Eb/No 0.31 db Margin 6.03 db This data rate is variable as the distance between Earth and Mars changes, with a range from 175 kbps to as high as 9 Mbps. These ranges are extreme and only appear in corner cases when Earth and Mars are at their farthest and closest distances from each other. A typical data rate for this segment will be approximately 550 kbps per orbiter. More than one orbiter may be able to communicate back to Earth simultaneously during this phase, depending on the availability of additional antennae on the Deep Space Network ) Satellite Bus Selection Requirements were defined to select the best possible commercial bus for the SmallSats. The derivations of these requirements were discussed in detail in section and are summarized in Table 5. Illinois Space Society 15

17 Table 5: SmallSat Hardware Requirements Specification Required Value Payload Power 177 W Payload Mass 38.2 kg Payload Form Factor Be able to support a 1.5 m antenna as well as two, smaller low gain antennas, all on separate faces Lifetime 7 years Several off the shelf small satellite busses were considered to make up the orbiting constellation of the MOSAIC program. Through this trade study summarized in Table 6, the TerraSense 300 small satellite platform manufactured by Dynetics was determined to be the optimal bus for carrying out the functions of the small communications satellite. According to the estimations of the payload, the payload mass capability had to be at least 38.2 kg with a payload power of at least 177 W. The TerraSense 300 meets this mass requirement, however none of the stock busses studied meet this requirement for Mars orbit. The required power in LEO for the SmallSats would be 412 W due to the 43% decrease in power production due to decreased solar flux. The stock TerraSense 300 will need to be modified to include extend the solar arrays from 260 W to at least 412 W in LEO. This extension will achieve sufficient power to support the communications payloads, and will require a minimal increase in spacecraft mass and cost. Even though the total mass of the bus structure is larger than compared to the other options, the other desirable traits of the TerraSense 300 bus make it the best option. The benefits in mission lifetime, payload data, versatility, and pointing accuracy set the TerraSense 300 apart from other off the shelf options. Criteria Table 6. Trade Study on the Small Satellite Bus Structures Ball BCP 100 [16] Surrey SSTL-300 [17] Dynetics TerraSense 180 [18] Dynetics TerraSense 300 [18] Orbital ATK A150 [19] Millennium Space Systems Altair [20] Max Payload 70 kg 150 kg 50 kg 110 kg 60 kg 50 kg Mass Total Mass 180 kg 368 kg 180 kg 300 kg 200 kg Payload Power (Earth) Payload Data Mission Lifetime Pointing Accuracy 200 W 280 W 200 W 260 W 150 W 90 W 2 Gb 16 Gb 8 Gb 1 TB 71% at 5 years 7 years 5-7 years 5-7 years 5 years 0.03º 0.1º 0.1º 3-axis 0.1º 3-axis 0.05º 0.006º Illinois Space Society 16

18 The mission lifetime of the satellite will be important for this communications network to minimize the frequency of replacing the small satellites and consequently minimize launch costs. The Monte Carlo analysis described later in this section imposed a 7 year lifetime requirement for the satellite. According to a datasheet of the TerraSense 300 released by Dynetics, the mission life of the satellite is 5-7 years [18]. Also, a datasheet released by OmniEarth, a satellite information services company that partnered with Dynetics to create a 15-satellite constellation, stated that the lifetime of a modified version of the TerraSense 300 would be 7-10 years [21]. On average, TerraSense 300 is designed to survive longer than other small satellite buses. However, these lifetime approximations were made assuming that the satellites will be in LEO. The lifetime in Mars orbit will be much lower because of the effects of greater radiation levels. For this communications network, the TerraSense 300 will be modified and improved to have a lifetime of approximately 7 years in Mars orbit. A model of the modified TerraSense 300 used for the MOSIAC architecture can be seen in Figure 9. Figure 9: A TerraSense 300 seen in orbit around Mars with its communications payloads. A simulation run in The Space Environment Information System (SPENVIS) program, a tool created by the European Space Agency, calculated that the total cumulative radiation for the SmallSat mission. This simulation assumed a 10,716 km orbit above mars with an inclination of 55 degrees. Taking into account both solar and galactic cosmic radiation, the Total Ionizing Dose (TID) experienced by a small satellite in Mars orbit after 7 years would be Rads [22]. This Illinois Space Society 17

19 figure is consistent with data collected from an experiment onboard the Mars Odyssey orbiter. [23] This is substantially a larger amount of radiation than a satellite would experience in LEO and will have a larger effect on the functions of the satellite. The Jet Propulsion Laboratory s design standards state that missions should be designed with a radiation design margin of 2 on a spacecraft level and 3 on a component level. [24] For the SmallSats in this mission, a value of 1,060 rads TID will be applied as requirement to all SmallSat components. Volume constraints of the payload fairing are also important to consider in any analysis. A rough mock-up of the SmallSats located inside of a 5 m short fairing (common to all Atlas V variants) has been created to show compliance with these size constraints. All four of the SmallSats along with the STAR 63F motor (the largest of the options for this architecture) have been included in Figure 10. Figure 10: TerraSense 300 satellites (black) arranged around the ESPA Grande ring (orange) with a STAR 63F (green) attached shown inside an Atlas V 5 m fairing. As a component for a communication constellation in Mars orbit with constant communication from the surface of mars to the relay satellite or directly back to earth, the satellite must also be able to store a large amount of data. The TerraSense 300 is overwhelmingly the best with the capability of storing 1 TB of data. One terabyte of data storage will be sufficient for the function of the small satellites but the hardware will be upgraded to have enhanced protection from additional radiation in Mars orbit rather than LEO. Illinois Space Society 18

20 Another important factor in choosing the satellite bus was the versatility of the design. For the purposes of this mission, the satellite bus had to be physically versatile to attach antennas on the sides. The TerraSense 300 datasheet states that it has a payload capacity on the nadir, ram, and wake sides. A parabolic antenna will be attached on the nadir side of the satellite facing either directly back to earth or to the large relay satellite depending on the phase of the mission. On the wake and ram sides, omnidirectional antennas will be symmetrically attached to receive data from the Mars surface. The two antennas will ensure that the satellite has coverage of the surface regardless of orientation. The pointing accuracy of the TerraSense 300 is sufficient to have limited effect on the downlink connection. The pointing accuracy of 0.1 causes at worst a loss of 0.30 db, which is relatively negligible considering other losses. The off the shelf version of the TerraSense 300 has a downlink of 1.2 Gbps using Ka-band from LEO. [18] This communication hardware will be adjusted to fit the requirements for this specific mission, however the high data rate proves that the satellite bus has the capability to process the data required by this mission (80 Mbps) with significant margin. Overall, it was determined that the TerraSense 300 by Dynetics would be the best small satellite bus for this communications network. Dynetics states that the TerraSense 300 can be launched as secondary payloads on a variety of rockets including the Falcon 9, Minotaur, and Antares. The satellite bus is also compatible with the ESPA Grande standard [25]. This standard was developed for Evolved Expendable Launch Vehicles (EELVs) including the Atlas V planned to launch these satellites. Up to four of these satellites can be fitted to one ESPA Grande ring, which is the configuration that will be used for the deployment of the MOSAIC satellites. [26] 4) Phase 2 High Altitude SmallSats and Laser Relay Satellite Phase 2 of the MOSAIC architecture sees the continuation of the small satellite constellation described in Phase 1, using identical SmallSats. These SmallSats cannot provide the data rates required to support a manned campaign to Mars. This phase also sees the addition of relay satellites beginning in 2031 at the Sun-Mars L1 point which allow for greater data transmission rates to support a human presence in the Martian system. A value of 80 Mbps was desired for this phase, a significant fraction of the 300 Mbps currently seen on the International Space Station. [27] This 300 Mbps figure was investigated by the team for feasibility but was ultimately determined too heavy and expensive of hardware to justify the investment. An 80 Mbps link will allow for significant amounts of data to be transferred, and still allow for a relatively lightweight architecture. 4.1) Orbital Mechanics 4.1.1) SmallSats An orbital constellation for the SmallSats has been chosen that utilizes two orbital planes, each featuring three satellites, to give full 100% coverage of the Martian surface. The satellites will all be placed in 55 degree inclination orbits, with the two orbital planes spaced apart by 90 degrees. While three satellites are sufficient for full coverage, a fourth is included in each orbital plane to allow for full in-space redundancy. In the nominal four satellite configuration, each satellite is separated by a true anomaly of 90 degrees. In the case of a failure, two of the remaining satellites will adjust their altitudes upwards to phase the remaining trio 120 degrees apart. Once this spacing is obtained, the satellites will return to their nominal orbital altitudes, providing complete coverage. Illinois Space Society 19

21 This configuration offers 100% coverage for the fewest number of satellites, while still being operationally sustainable. Other configurations that give 100% global coverage with as few as four satellites have been examined by Draim. [28] However, these constellations rely on highly elliptical orbits leading to heavily increased operational complexity and widely varying data rates. These traits are highly undesirable for a communications constellation. The configuration chosen relies on circular orbits, yielding relatively constant distance to ground targets. The minimum altitude above the Martian surface to achieve continuous 100% ground coverage was calculated via iteration in Systems Tool Kit (STK) and found to be 7,104 km. Due to the sloped nature of much of the Martian terrain, the constellation was designed to provide coverage for slopes up to 5 degrees. This 5 degree figure is consistent with the average slopes of several locations on Mars, including the well-known Olympus Mons. [29] This assumption raises the altitude of the satellites to 10,716 km to allow for more consistent coverage. Table 7. Table of Launch Vehicle and Insertion Stage Required for Various Launch Windows for the Phase 2 Altitude C3 (km 2 /s 2 ) Earth-Departure Insertion Stage STAR 31 Mars-Arrival STAR 48B STAR 48A STAR 63F Atlas V Used During phase two of the mission, the SmallSats are launched to the final communications orbit, which has an altitude of 11,706 km, instead of the 1,507 km orbit used in Phase 1. The Illinois Space Society 20

22 analysis for finding Mars transfer trajectories still applies, and the analysis for orbital insertion and phasing need only be slightly modified. For orbital insertion, the C3 requirements are the same for a given transfer trajectory, but the ΔV requirement for each trajectory changes with the altitude. This change will require a different insertion stage for some C3 values, which will necessitate a different launch vehicle in some cases. Table 7, shown above, is an updated launch vehicle trade space, with the next 9 launch windows beginning in 2031, using the same payload mass and an altitude of 11,706 km. The same 20 m/s phasing budget was used for the Phase 2 orbit, which resulted in a required phasing time of days for the two farther satellites, and a phasing time of 4.58 days for the two closer satellites. The phasing time required for the two farther satellites can be seen below in Figure 11. ΔV versus time required for phasing maneuver for Phase 2. These times, while longer than the times required for Phase 1, are still short enough, at less than two weeks, that they will not unreasonably delay the mission, as the spacecraft can conduct their final checkout and testing procedures during this period. Figure 11. ΔV versus time required for phasing maneuver for Phase ) Relay Satellites The relay satellites required to transmit at larger data rates from Mars are required to be kept at a point relatively close to the Mars orbiting SmallSats. This location must avoid eclipse as much as possible and require minimal energy to reach and maintain position at. The natural choice for such an orbit is the Sun-Mars L1 point. This point offers freedom from eclipse by the planet Mars Illinois Space Society 21

23 itself, requires relatively low energy to reach compared to Martian orbits, and has a very small station keeping requirement. [30] Locations such as the Sun-Mars L4/L5 points were considered as they offer freedom from solar eclipse, but the transmission distances from the SmallSats to the relay satellites were prohibitively large. The distance between Mars and the Sun-Mars L4 point is greater than the minimum distance between Earth and Mars, so the point has no utility as a primary communications relay. According to computations by Carrico, Strizzi, Kutrueb and Damphousse [31], the orbit insertion ΔV for Sun-Mars L1 for a 200 day transfer varies from km/s for the minimum energy method. Slower transfers will result in lower insertion energies, so the worst case assume ΔV requirement has been set to 1,700 m/s. The C3 energies required for launch vehicles for the windows studied in the paper range from km 2 /s ) Satellite Hardware 4.2.1) Communications Payloads In Phase 2, communications from the ground are assumed to come from a manned base on the surface of Mars, transmitting via a UHF link. This manned base is assumed to have large power capabilities (on the order of kilowatts), and an antenna with a diameter of at least 3 meters. This power level and volume constraint are well within the designs of most Mars mission concepts. [32] The SmallSats will still receive data from the ground in Phase 2, albeit at a greater data rate. Using the same SmallSat communications hardware from Phase 1 alongside new Mars surface assets, data rates of at least 80 Mbps can be reached as shown in the link budget analysis of Table 8. Rovers during this phase will require a greater transmitting power, but advances made in the intervening 20 year period have been assumed to be sufficient to make up this difference. Table 8: Downlink Budget for a Manned Base to the SmallSats Specification Transmitter RF Power Output Frequency Ground Antenna Gain Miscellaneous Rover Losses Polarization Loss Range to Spacecraft Path Loss SmallSat Pointing Error SmallSat Pointing Loss SmallSat Antenna Receiving Gain SmallSat System Noise Temperature Demodulator Implementation Loss Data Rate Attained Eb/No Required Eb/No Margin Value 600 Watts MHz dbic db db 13,405 km db 0.1 degrees db 6.77 dbic 526 K db 80 Mbps 6.60 db 0.31 db 6.29 db Illinois Space Society 22

24 Communications from the surface of Mars will be routed to the Sun-Mars L1 relay satellite in Phase 2 via an X-band communications link. The same 1.5 m antenna hardware used for the directto-earth link in Phase 1 will instead be pointed at the relay satellite at the Lagrange point. This new data link is outlined in Table 9. A high gain antenna with a diameter of 3 m has been chosen for the L1 relay satellite. This antenna will be nearly identical to the high gain antenna on the Mars Reconnaissance Orbiter, lending it significant successful flight heritage. The high gain antenna must be pointed at the actively transmitting satellite. Each of the two relay satellites can communicate with one of the SmallSats at a data rate of 80 Mbps, leading to a total maximum downlink for the system of 160 Mbps. The relay satellites must be pointed directly at the transmitting SmallSat, as attempting to cover all of the satellites by pointing directly at the center of Mars would result in an off angle of 0.58 in the worst case. This pointing inaccuracy would lead to a 6 db loss of gain on the 3 m receiving antenna, resulting a maximum data rate of 20 Mbps. With the high pointing accuracy of the relay satellite required for laser communications, pointing losses from this spacecraft are negligible and are not included in this analysis. A transponder produced by Thales Alenia Space has been chosen to provide data. The X/X/Ka Deep Space Transponder provides the ability to uplink and downlink high data rates through the relay satellite. The Ka-band hardware included also has the benefit of forwards compatibility with other Martian systems. This transponder operates at a maximum power of 32 Watts, and has a mass of only 3.7 kg. Table 9: Downlink Budget for SmallSats to the Sun-Mars L1 Relay Satellite Specification Value Transmitter RF Power Output 100 Watts Frequency 8,400 MHz SmallSat Antenna Gain dbic Miscellaneous Rover Losses db Polarization Loss db Range to Spacecraft 1,050,000 km Path Loss db SmallSat Pointing Error 0.1 degrees SmallSat Pointing Loss db Relay Satellite Antenna Receiving Gain 48 dbic Relay Satellite System Noise Temperature 60 K [33] Demodulator Implementation Loss db Data Rate 80 Mbps Attained Eb/No 6.45 db Required Eb/No 0.31 db Margin 6.14 db The relay satellites in the MOSAIC architecture transmit directly from the Sun-Mars L1 point to the Earth using a laser communication terminal. Laser (or optical) communication was chosen for this link due to the unprecedented data rates that are required of a manned Mars architecture. Illinois Space Society 23

25 Optical communication offers significant benefits in power consumption and transmitter form factor over traditional RF designs. When the sun is directly in between the Earth and Mars, laser communication also has the advantage of allowing transmission for a longer period of time. The so called solar exclusion angle or point at which the Sun blocks all direct communication between two points is as low as 3 for laser communication. [34] This angle is the angle between the Sun and Earth from the perspective of the spacecraft, known commonly as the Sun-Earth Probe (SEP) angle. For an X-band communication link, this angle is approximately 5, giving laser communication a distinct advantage. The blackout period for laser communication is approximately 21.5 days, while X-band has a coverage gap of about 35.9 days. The laser communication system chosen for the relay satellites features a transmitter with an input power of 2 kw. This is significantly greater than any tested system to date, as the current state of the art LADEE mission used a laser with an input power of 137 W. [35] The data rate was calculated using the following equation obtained from [36]: R = P tτ opt τ ATM A π(θ t 2) 2 L 2 E p N b Where Pt is transmitter power, τopt is optical efficiency, τatm is the atmospheric attenuation factor, A is receiver telescope area, θt is transmitter divergence, L is path length, Ep is photon energy, and Nb is the number of photons per bit. To ensure the accuracy of the equation, values were obtained from different sources of comparable optical systems with known data outputs. The values were put into the MATLAB code and the data rates received as output were similar, validating the model. The values used for this equation in the primary downlink calculation from LACOR to the Earth are shown in Table 10. The transmitter divergence was a limiting factor in both the data rate and required pointing accuracy for the satellite bus on which the optical system is located. The smaller the transmitter beam divergence, the higher data rate is possible but the pointing accuracy also becomes much higher. The optimal beam divergence for the balance between data rate and pointing accuracy was found to be 16 x 10-6 radians through changing the values of the transmitter divergence in the code for data rate and pointing accuracy until an acceptable balance was met. This value was proven feasible by a successful NASA test of an optical transmitter with an equal beam divergence. [37] The highest data rate achieved by the system at the closest distance to Earth (.36 AU) is 3,734.4 Mbps. The lowest data rate, seen at the farthest distance from Earth (2.68 AU) is 69.6 Mbps. The data rate at the average distance to Earth (1.52 AU) is Mbps. A plot of the data rate as a function of distance slopes steeply as the satellite nears its closest distance to Earth so high data rates can be expected during that period. The equation used to calculate pointing accuracy in arcseconds is derived from the geometry of the beam over the Earth-Mars distance: Pointing Accuracy = ( θt ) 206, For the system implemented a continuous pointing accuracy of at least 1.65 arcseconds was identified as the requirement for the laser communication satellite. Illinois Space Society 24

26 Table 10: Laser Communication Data Rate Calculation Values from the L1 Relay to Earth Specification Value Transmitter Power 2,000 Watts Wavelength 1,550 x 10-9 m Optical Efficiency 0.43 db Atmospheric Transmission db Transmitter Divergence 16 x 10-6 rad Diameter of Receiver Telescope 10 m Receiver Sensitivity 100 photons/bit Given the two large separate sets of communications systems and much higher payload requirements, the relay satellite has larger payload mass, power, and volume constraints than the SmallSats. The mass breakdown for these satellites is summarized in Table 11. The mass of the laser communication terminals was estimated as a direct scaling from the Laser Communications Relay Demonstration. [38] Table 11: LACOR Communication Payload Mass Budget Component Total Mass [kg] High Gain Antenna [6] 18.9 X-band Transponder [39] 3.7 Laser Communication Gimballed Telescope [38] 80.0 Laser Communication Modem/Electronics 40.0 Misc. Wiring 4.0 Subtotal Contingency (43%) 63.0 Total This total mass of kg, along with the 2,035 W total draw of the communication system are some of the driving constraints for the satellite bus selection ) Satellite Bus Selection Based on the proposed constellation configuration and current technology levels for relevant satellite and communication systems, minimum requirements for the satellite hardware were defined. The derivations of these requirements were discussed in detail in section and are summarized in the table below. Table 12: LACOR Hardware Requirements Specification Required Value Optical Transmitter Power 2,000 W X-Band Transmitter Power 35 W Bus ΔV Capability 1,730 m/s Pointing Accuracy 1.65 arcseconds Payload Mass kg Lifetime 15 years Illinois Space Society 25

27 The satellite bus used for the Laser Communications Relay Satellite (LACOR) is the Orbital ATK GEOStar-2 Bus. The GEOStar-2 s low cost of $150 million and customer confidence make it a clear contender amongst similarly sized and capable satellite buses currently in production including the SSL 1300 and I-3K. [40] [41] A full trade study of the SSL 1300 and the GEOStar- 2 was created and can be seen in Table 13 below. The GEOStar-2 was ultimately chosen as it meets all mission requirements within smaller mass and volume constraints. Table 13: Relay Satellite Bus Trade Study Criteria SSL 1300 GEOStar-2 Manufacturer Space Systems Loral Orbital ATK Payload Power (Earth) 5-25 kw [42] 5.5 kw [43] Launch Mass 5,500 kg [44] 3,325 kg Max Payload Mass 1,500 kg [45] 500 kg Size Can fit in 5 m Fairing 1.75 m x 1.7 m x 1.8 m [46] Cost $200 million $150 million [40] Previous Relevant Missions GEO GEO [41] Lifetime 15+ years [47] 15+ years [41] Current Pointing Accuracy Pitch degrees Yaw degrees [48] Pointing Control degrees Standard Pointing Knowledge degrees [49] Modifications Needed Better Pointing Accuracy Needed Better Pointing Accuracy Needed The GEOStar-2 s 25+ mission history presents clear competency, and the 15+ year typical mission lifetime, and lighter comparative launch mass makes the GEOStar-2 the decisive option. [41] [50] It also features radiation hardened processors certified for deep space use. [51] It is critical for keeping a strict production timeline that as few changes to the bus are made as possible to meet mission criteria. The radiation hardened certified processors constitutes one less modification needed to integrate a commercially available communications satellite with the necessary deep space communications equipment. Two satellites will be sent to the Sun-Mars L1 point to provide full redundancy for the high speed communications link. A model of one of the LACOR satellites can be seen in Figure 12. Illinois Space Society 26

28 Figure 12: Model of LACOR satellite with laser communication terminal shown on the top of the bus and the X-band antenna on the bottom. The GEOStar-2 features a 500 kg payload for a total launch mass of 3,325 kg. [41] The low payload and launch mass are desirable to keep launch costs to a minimum and the necessary payloads fit into the weight restriction with substantial margin. The bus has an estimated lifetime of greater than 15 years in GEO, a relatively similar environment to that found at the Sun-Mars L1 point. The low launch mass, and size of only 1.75 x 1.7 x 1.8m, gives flexibility in choosing launch vehicles. Past launch vehicles for the GEOStar-2 include: Falcon 9, Ariane, Proton, and others. [41] In LEO, the bus will provide 5,500 W of power. [41] Near Mars orbit, the power provided by the stock bus will be reduced to 43% (due to the decrease in solar intensity), producing only 2,365 W. [52] This still meets the 2,035 W requirement for the communications payload (as determined in section 4.2.2) and upgraded attitude determination system discussed below. A minimum pointing accuracy of arcseconds was determined for optical communication between Sun-Mars L1 and Earth. The GEOStar-2 bus in its stock configuration only provides a pointing control of 36 arcseconds and standard pointing knowledge of 108 arcseconds. [53] Augmenting the GEOStar-2 s attitude determination and control system is therefore necessary. The only viable system to achieve the necessary pointing knowledge is Ball Aerospace s High Accuracy Star Trackers (HAST), which can individually achieve pointing knowledge of arcseconds. [54]. To further augment this system, two HAST will be placed with perpendicular boresights, one out of the plane of the solar system; this configuration ensures maximum pointing knowledge by providing accurate orientation information for multiple, orthogonal degrees of freedom. The HAST weighs less than 6.8 kg and consumes no more than 120 W. [54] The GEOStar-2 attitude control system uses reaction wheels and may require an upgrade from the stock condition of the significantly increased pointing knowledge does not reduce the pointing control sufficiently. This is a standard upgrade Orbital ATK provides and will not impact the benefit of using a stock bus. [31] Illinois Space Society 27

29 To operate as a relay satellite, LACOR must have an X-Band antenna facing Mars, opposite of the GEOStar-2 s standard nadir-pointing antenna configuration in addition to the optical Earthpointing antenna. [41] These changes will constitute the greatest modifications to the stock bus configuration, but should not require any extensive rework of the bus structure. For standard operation, LACOR must have sufficient fuel and capability to complete its orbit insertion maneuvers and maintain its orbit for a minimum of its expected lifetime. The stock bus has a total ΔV capability of 1,763 m/s, typically used for transfer from a Geostationary Transfer Orbit into a Geostationary Orbit. The required ΔV for LACOR orbit insertion is 1,700 m/s, so 63 m/s is left over to provide for station keeping at the L1 point. The stationkeeping requirements for this point are expected to be as low as 2 m/s per year, providing sufficient propellant to maintain positioning for the 15 year nominal lifetime with significant margin. [30] The launch configuration inside of an Atlas V 5 m payload fairing is shown in Figure 13. This launch vehicle has been used to launch satellites of the GEOStar series in the past so no integration issues are anticipated. Figure 13: LACOR shown in the launch configuration with the high gain antenna located at the top of the diagram. 4.3) Production Schedule and Launch Frequency 4.3.1) SmallSat Launch Analysis A Monte Carlo analysis was performed to determine the rate at which satellites in the constellation must be replaced. Satellite lifetime was assumed to be a normal distribution. The Illinois Space Society 28

30 mean SmallSat lifetime was placed at 7 years (1 year transit + 6 year operation), with a standard deviation of 1.5 years. [55] The constellation must have at least six active satellites to allow for 100% coverage of the entire Martian surface, so only runs in which this value was maintained over the 30 year period reaching from 2035 to 2065 were considered to be successes. Runs which dropped below this level even momentarily were ruled failures. A success rate of at least 98% was desired to define an architecture option as viable. The results from this study led the team to decide that a six year build-up would be required to allow the constellation to be fully active, leading to the first launches to this high orbit in The analysis also showed that a four satellite launch should occur every synodic period (approximately every 780 days). A Monte Carlo simulation with 1,000 independent runs was performed, and the probability of maintaining this coverage level over the 30 year period of study was found to be 98.4%. Over all runs (30,000 total years studied), six satellite coverage was shown to be maintained 99.99% of the time using this configuration. An example of a successful run where full six satellite coverage was maintained from 2035 to 2065 is shown in Figure 14. This analysis gives the team confidence that the chosen architecture will be sufficient to maintain continuous coverage of the Martian surface by the satellites. Figure 14: Plot showing a successful example run of the lifetime analysis. The dotted red line shows the required number of satellites to achieve 100% coverage. The MOSAIC architecture calls for the manufacture of 40 SmallSats within the first few years of inception. These satellites will be launched four at a time over a 22 year period, but the bulk purchase enable a lower per unit cost. The satellites not launched immediately will be stored on the ground, and fueled just prior to launch. Past satellites have had long ground wait times, such Illinois Space Society 29

31 as the DSCOVR satellite which launched in February of This satellite completed construction in 2001, but was grounded for flight for 14 years due to political turnover. [56] The total launch schedule for the entire MOSAIC architecture can be found in tabular form in Appendix E ) LACOR Launch Frequency These satellites have a nominal lifetime of 15 years, after which a new set of satellites will be sent. If the original set of satellites is still active, they can serve as on-orbit spares or as additional links in the communications chain. The GEOStar-2 production time is 24 months. [41] With its required modifications, an additional six months are added to allow for integration and testing of the attitude control system. The LACOR production timeline is 30 month lead. 5) Phase 3 Solar Eclipse Relay Satellite Phases 1 and 2 form a build-up of capability to allow for 100% coverage of the Martian surface during most of the Earth-Mars cycle. However, these satellites will be obstructed by the Sun once every synodic period for 16.4 days, resulting in a communications blackout. In Phase 3 the Solar Eclipse Relay satellite (SER) will be sent to the Sun-Earth L4 point to provide a relay for the MOSAIC architecture when the Sun is directly between the Earth and Mars. This relay will function with a laser communication receiver and transmitter, and will pass data around the Sun to and from Earth, albeit at a lower data rate than during the non-eclipsed portions of the mission. The majority of Phase 3 will see the full data rates of Phase 2 through an identical architecture. This satellite simply adds a capability required to achieve the minimum coverage figure of 98% year-round. 5.1) Orbital Mechanics According to the analysis performed by Llanos, Miller, and Hintz, the orbit insertion energy for transfer to Sun-Earth L4, if launched in October, was found to be 6.76 km 2 /s 2 to enter the transfer orbit, and km/s to enter the parked, halo orbit. [57] The time of flight was calculated as days. While this transfer is quite long, the relay satellite will be able to serve in its function most of the time while still on the trajectory to the point. This launch is required to occur in October, because the position of the Earth on its orbit around the Sun influences the optimality of the trajectory to the L4 point. Accounting for this change, July was selected as the time for launch, since it had preferable requirements for orbit insertion energy and C3 energy. The Sun-Earth L4 point was selected as a destination in order to account for gaps in coverage, caused by Earth being eclipsed by the Sun. During the eclipse the Earth will be unable to receive any signals from both Mars and Sun-Mars L4, since both of them will be hidden by the Sun. The solution proposed to this, would be to place a relay satellite at either the Sun-Earth L4 or L5, which are, in this case, functionally equivalent. The Sun-Earth L4 point was picked because the orbit insertion energy, the C3 energy, and time of flight were preferable for the data available. 5.2) Satellite Hardware 5.2.1) Communications Hardware The communications hardware for the Solar Eclipse Relay satellite will be identical to some of the hardware included on the LACOR satellite in Phase 2. Instead of including an X-band antenna and a laser communications terminal, two laser communications terminals are used. The first of these terminals is identical to those onboard LACOR and will point towards Earth. The second laser terminal will have a diameter of 3 m, giving it the same form factor as the LACOR Illinois Space Society 30

32 satellites. This larger diameter will allow for significantly larger data rates during blackout periods. Each of these lasers features the same 2,000 W power draw as those found on the LACOR satellite. This allows the second, larger diameter telescope to share common hardware with the other systems. A mass breakdown of this payload is included in Table 14. Table 14: SER Communications Payload Mass Budget Component Total Mass [kg] Laser Communication Gimballed Telescope (1 m) 80.0 Laser Communication Gimballed Telescope (3 m) Laser Communication Modem/Electronics (2) 80.0 Misc. Wiring 4.0 Subtotal Contingency (43%) Total Links between the ground, SmallSats, and the LACOR satellites are identical to those found in section However, instead of communicating directly back to Earth, the LACOR satellite will instead point towards the SER satellite while eclipsed by the Sun. Calculations to determine the communication data rates available were performed using the same MATLAB code described in section The downlink portion of this phase has two segments, one from LACOR to SER, and one from SER to the ground. The maximum data rate found for the LACOR to SER link was 17.4 Mbps, while the SER to ground phase permitted a data rate of Mbps. In this case the satellite to satellite link is the weakest, and the 17.4 Mbps data rate obtained is a significant drop from the 69.6 Mbps minimum attained without the relay. However, this communications link will be invaluable to any human installation on Mars, permitting continuous monitoring and scientific return. A similar analysis was performed for uplink, yielding a data rate of 1.9 Mbps. This link is limited by the ability of the ground based laser terminals on Earth to communicate up to the L4 point, with expected uplink rates on the order of 2 Mbps. [58] [59] Earth to SER link is discussed in greater detail in section 6 of this report. A maximum uplink rate of 1.9 Mbps was therefore assumed for this architecture ) Satellite Bus The requirements identified for SER in the sections above are summarized in Table 15. These requirements are similar to those of LACOR, but with increased mass and power constraints. Table 15: SER Hardware Requirements Specification Required Value Optical Transmitter Power (2) 4,000 W Bus ΔV Capability 740 m/s Pointing Accuracy 1.65 arcseconds Payload Mass kg Lifetime 15 years A GEOStar-2 bus has also been chosen for the LACOR satellite, as the stock GEOStar-2 satisfies all of the requirements with the exception of payload mass. The payload mass for SER is Illinois Space Society 31

33 greater than that of LACOR due to the duplication of the heavy optical components. Fortunately, the insertion into the Sun-Earth L4 point requires only a fraction of the ΔV capability of the GEOStar-2 bus (710 m/s vs. 1,763 m/s), so the bus can be underfueled by several hundred kilograms to more than make up for this difference. The power system for SER will be able to provide 5,550 W, since the satellite will be located at 1 AU from Earth orbit. This power production capability meets the requirement set by the communications payload. Identical modifications to those made on LACOR will be required for the attitude determination and control system, with higher accuracy star trackers allowing for more precise pointing of the laser terminals. 6) Ground Infrastructure and Operations The direct to Earth X-band communication during Phase 1 of the MOSAIC architecture will utilize the Deep Space Network 34 m antennae. This ground infrastructure is assumed to be maintained at current capacity through at least 2032 for the purposes of this study. In order to complete the necessary infrastructure for an optical communications link, an Earthbased optical terminal will have to be constructed. This terminal will need to be capable of receiving and transmitting at appropriate data rates, while also maintaining the required 98% availability. Two options were initially considered for such a terminal: a single orbiting relay satellite or a ground-based network of multiple stations. With a satellite in orbit around the Earth, the primary benefit is the avoidance of any atmospheric interference with the optical signal. This leads to very high availabilities, with one particular study demonstrating 98% availability for an Earth-orbiting satellite with a 7 m aperture telescope. The design of this particular satellite was based on the Next Generation Space Telescope, a project which later became the James Webb Space Telescope. Still, despite the high availabilities that a satellite would offer, the system also has key disadvantages. A single satellite orbiting Earth can only track and communicate with one spacecraft at a time, unlike a more desirable ground-based network that could track multiple targets if needed. [60] Beyond that, any kind of space-based system would also be tremendously expensive. The James Webb Space Telescope, the basis for the design of the satellite mentioned earlier, is currently projected to cost $7.998 billion. [61] Thus, with both operational and budgetary considerations in mind, it was decided to focus on a series of ground stations for the Earth-based optical terminal. 6.1) Site Locations for Optical Ground Network Several parameters must be considered when selecting site locations for a reliable groundbased optical communications network. These parameters primarily consist of restrictions on where individual sites can be located, as determined by criteria for the overall availability of the network. One of the first requirements to consider with any Earth-to-space communications network is its line of sight (LOS) coverage. Optical ground stations must be spread around the globe in a configuration that ensures there is always at least one station with a direct line of sight to the Sun- Mars L1 relay satellite. This is accomplished by choosing sites relatively close to the equator, while also spacing stations as evenly as possible around the Earth. An ideal network is able to maintain continuous LOS coverage by passing a signal between stations, allowing one station to take over for another when the latter loses LOS due to the Earth s rotation. In addition to LOS coverage, the network s weather availability also plays a major role in determining individual site Illinois Space Society 32

34 locations. Optical communications performance is heavily dependent on atmospheric conditions, and signals degrade severely with the presence of any opaque cloud cover. It is thus vital that ground stations are located in dry, typically sunny regions where the threat of cloud cover can be kept to a minimum. The distance between stations should also be large enough to ensure that there is no correlation in weather patterns between sites. Finally, the network greatly benefits from keeping individual stations at high altitudes, where the thinner atmosphere helps reduce signal degradation. [62] In choosing a final layout for the ground network, the MOSAIC team looked at two options previously investigated by the Jet Propulsion Laboratory: a linearly dispersed optical subnet (LDOS) with eight stations or a 3 x 4 clustered optical subnet (COS) with twelve stations. The LDOS configuration would place eight stations more or less evenly around the Earth, while the COS configuration would contain four clusters of three stations. Both strategies are expected to have LOS coverage of 100%, however they do diverge when determining their respective weather availabilities. A COS 3 x 4 has a weather availability of 96% when the system transmits at a maximum zenith angle of 60 degrees. An LDOS with 8 stations can also achieve 96% availability, although this figure is calculated using a maximum zenith angle of 75 degrees. [62] Still, despite the COS 3 x 4 design offering similar performance at a lower zenith angle, the LDOS with eight stations became the ground network of choice for MOSAIC. The reasons behind this decision are both budgetary and geographical. The LDOS requires four fewer sites than the COS 3 x 4 design, leading to significant reductions in the cost of the network. In addition, the COS 3 x 4 requires that one of its four clusters be located in Pakistan, where thus far studies have only identified one of the required three station sites. The LDOS with 8 stations only requires the singular site in Pakistan and hence does not have this problem. [62] A worldwide map of stations for the LDOS network can be seen in Figure 15, with details on individual site locations available in Table 16. Lastly, it should be noted that although the optical ground network will only have 96% availability, X-band transmissions will also be possible at low data rates between SmallSats and the existing Deep Space Network. This feature can be used to fill in any availability gaps in the optical network if communication to Earth is urgently needed. Figure 15: Mapped site locations for an LDOS with eight ground stations (world map background from [63]) Illinois Space Society 33

35 Table 16: Detailed Site Locations for an Eight-Station LDOS (reproduced from [62]) Location Altitude [km] Coordinates [deg] Cloud-free days/weather Preexisting infrastructure Table Mountain Facility, California N, 118 W 66%/arid Yes Mauna Kea, Hawaii N, 155 W >69%/dry Yes Siding Spring Mountain, Australia S, 149 E 67%/dry Yes Mt. Bruce, Australia S, 118 E NA/dry Information N/A Ziarat, Pakistan N, 68 E 69%/arid Information N/A Jabal Ibrahim, Saudi Arabia N, 41 E NA Information N/A Calar Alto, Spain N, 2 W 67%/arid Yes Cerro Pachan, Chile S, 71 W 77%/arid Yes 6.2) Technical Specifications for Ground Telescopes As determined by the optical link equation, each ground station will require a 10 m diameter telescope to receive transmissions from the Sun-Mars L1 relay satellite. A separate, smaller telescope equipped with a high-power laser system will also be located at each station to provide uplink capability. Both the uplink and downlink systems will be enclosed in protective dome-like structures, similar to those seen on current ground-based telescopes. The downlink system for each station will be based on a design for a 10 m optical receiver telescope that was proposed by the Jet Propulsion Laboratory in This design employs a 10 m-diameter segmented primary mirror, with the focal length also kept at 10 m to minimize the telescope s overall size and cost. The primary mirror will be constructed from lightweight glass, and every mirror exposed to the elements will be equipped with a CO2 snow cleaning system to eliminate airborne contaminants. JPL s research also compared different telescope configurations, namely prime-focus and Cassegrain designs. Although the study chose not to specify a superior design, the MOSAIC team believes that a Cassegrain configuration is the best choice. Although initially a Cassegrain-style telescope presents additional manufacturing difficulties, in the long term it is easier to operate and maintain. The configuration uses a 10 m spherical primary mirror to concentrate the incoming signal onto a smaller, secondary mirror. The aspherical secondary mirror then reflects the signal back to the optical detector through a hole in the primary mirror. Figure 16 provides a diagram of this basic Cassegrain design. Not pictured are two small corrective optic elements which are placed near the focal plane. Once the signal passes through the telescope s primary optics, it will enter the detector array shown in Figure 17. The incoming signal is reflected off two small mirrors and then passed through a bandpass filter, designed to block any stray energy not at the correct frequency. At the end of the process is the communication detector, which receives and interprets the optical signal. The communication detector is also surrounded by a pointing detector, which registers signal fluctuations and directs the telescope to move Illinois Space Society 34

36 accordingly to correct. Movement of the telescope will be via a conventional mount, similar in style to what is presently used at the Keck Observatory in Hawaii. The required tracking accuracy of the mount will be achieved using technology currently in use by microwave antennas in the Deep Space Network. These contemporary DSN systems are capable of achieving the required pointing accuracy of ±50 µrad. [64] Figure 16: Cassegrain configuration for 10 m optical receiving telescopes (diagram from [64]) Figure 17: Optical signal detection system for 10 m receiving telescopes (diagram from [64]) Illinois Space Society 35

37 The structure of the uplink telescope will be based on the NASA/JPL Optical Communications Telescope Laboratory (OCTL), with upgrades to internal components to achieve desirable data rates. Like the OCTL, each uplink telescope will utilize a Coude mirror configuration with an aperture 1 m in diameter. [65] The system will also contain a high-power, multi-beam laser system to produce the optical signal and act as a tracking beacon for the Sun-Mars L1 relay satellite. At present, high-power laser systems developed for the Department of Defense could enable kilobit per second uplink capabilities to targets in the Solar System. [66] NASA s stated goal is an uplink rate of 2 Mbps, so additional investments will be made to develop a laser generation system with sufficient power. [65] Uplink performance will also be improved by dividing the laser s total power between multiple beams, a technique which helps reduce signal degradation due to atmospheric interference. Data from the Ground/Orbiter Lasercomm Demonstration (GOLD) has demonstrated that multi-beam transmission helps avoid signal fluctuations, improving the overall quality of the transmission. [67] The uplink telescope s anti-contamination instruments, mount, and pointing system will be similar to those discussed for the downlink system. Finally, although the protective dome structure will offer protection in case of severe weather, in general all ground stations will operate day and night to provide complete coverage. To prevent damage when operating at low Sun-Earth-Probe angles in the daytime, every telescope dome will be equipped with an inflatable sunshield made from aluminized Mylar. This design is costeffective and easy to modify, with the added bonus of isolating the sunshield from the telescope mount. This prevents the optics from being disturbed if, for example, the sunshield encounters any wind. In addition, temperature control systems within the domes will counteract any unwanted heat buildup due to solar radiation. [64] 7) Programmatic Considerations 7.1) Architecture Cost Analysis Costing for the MOSAIC architecture was performed by finding cost analogs for all major components of the architecture. The main cost goal of the program besides the obvious goal of minimizing total cost, was to spread out cost over all years to allow the architecture to fit within the current budget-constrained environment faced by NASA. All figures shown in the below section in 2015 dollars. For the SmallSats, a costing methodology was used that takes into account the large volume of satellites that will be purchased. The theoretical first unit cost was estimated using the SSTL- 300 satellite commercial competitor to the TerraSense 300, as no data on the TerraSense itself was found. This analog satellite also conforms to the ESPA Grande standard and has similar total mass, payload mass, and payload power. The purchase price of one of these satellites is listed to be $26.09 M. [68] The calculation of the total SmallSat production cost used the equation Production Cost = TFU N 1 ln ( S ) (2) ln (2) where TFU is the theoretical first unit, N is the number of units to produce, and S is the Learning Curve Slope was used to calculate the total production cost of a run of 40 satellites. [69] These 40 satellites will launch in the windows from 2022 to 2046, at which point the SmallSat architecture will be reassessed. The theoretical first unit cost used is the $26.09 million figure from the SSTL-300 spacecraft bus. For an N of 40 satellites, a learning curve slope of 90% is applied for aerospace hardware. This analysis brings the total cost of this 40 small satellite production run to $596 million. This 100% Illinois Space Society 36

38 cost was spread over a five year period from 2018 to 2022, in which the actual production will likely occur. The average cost per satellite for this run is $14.89 million, a steep discount from the TFU. This cost learning curve was not applied to the large relay satellites, as at most two identical ones will be produced at any given point. The cost of the relay satellites has been estimated to be $150 million per unit, equal to the Thaicom 8 satellite which also utilizes the GEOStar-2 bus. [40] The cost of the communications modifications to the bus is included in this figure. For the Atlas V launch vehicles, an average cost of $185 million per launch was assumed as part of a bulk buy that will see a launch every two years on average. This figure is based on the Department of Defense reported expenses on launch vehicles for [70] Most of these launches will utilize an Orbital ATK STAR motor variant. For these launches an additional cost of $8 million has been added for the purchase and integration of the STAR motor. This figure comes from the launch cost of the Minotaur IV, which uses a STAR 48 variant as its optional fourth stage. [71] [72] The estimated budget for the optical ground network is based on the construction costs of preexisting 10 m and 1 m telescopes. Additional investments will occur in high-power uplink lasers and the necessary infrastructure for undeveloped station sites. First, the estimated cost of the eight receiving telescopes is considered. The 10 m optical receiver telescopes have a primary mirror and mount system that is very similar to components in the twin 10 m Keck Telescopes in Hawaii. Each Keck Telescope cost $96.15 million to develop and requires $15.4 million to operate annually, with instrumentation costs excluded. [73] Considering the additional development costs associated with creating suitable optical instruments, instrumentation costs per telescope are anticipated to be higher than Keck. An initial development effort to research and produce suitable optical instruments is being given a $100 million budget, with post-development instrumentation costs estimated to drop to $50 million for each of the remaining seven telescopes. Multiplying the development costs of Keck and factoring in the required instrumentation investment, the combined cost of all eight 10 m optical receiver stations is estimated to be $1.22 billion with annual operating costs of $123.2 million. Next, turning to the 1 m optical transmitter stations, a cost evaluation was performed for researching, developing, and producing eight high-power laser generation systems. A contemporary example is the 30 kw Laser Weapon System, developed by the Office of Naval Research at a cost of $40 million dollars. [74] That said, the MOSAIC team anticipates that the high-power lasers required for optical transmission at 100 kbps will need to be even more powerful than the current state-of-the-art. $100 million will be allocated to the development of such a system, with the estimation that the cost-per-unit will decrease to $25 million once development is complete. The price for each 1 m telescope and observatory is based on a 2004 feasibility study for building a similarly-sized telescope in the state of North Dakota, a report which quoted a price of $1.9 million for such an observatory. [75] The actual cost for each optical transmitter telescope is estimated to be at least $5 million, over double the North Dakota figure, to account for the additional pointing and tracking systems that need to be included. Factoring in the development costs of a suitable high-power laser system, as well as the price of each 1 m observatory, the cost for the eight optical transmitter stations is projected to be $315 million. Operating costs are estimated at $3 million per telescope per year, yielding a total operating cost of $24 million annually for the transmitter stations. Illinois Space Society 37

39 Lastly, the team considered the costs associated with building at the three ground sites with minimal preexisting infrastructure for an observatory: Mt. Bruce in Australia, Ziarat in Pakistan, and Jabal Ibrahim in Saudi Arabia. $25 million dollars will be dedicated solely to infrastructure development at each site, with a focus on developing sufficient power generation systems and a road network for transporting construction materials. Taking into account infrastructure investment as well as the previously-calculated costs of the ground stations themselves, the grand total for the development cost of the optical ground network is $ billion. Operating costs will total $147.2 million per year. The total cost of the MOSIAC architecture over the timeframe of the architecture comes to $14.25 billion, with annual costs ranging from $200 million to $500 million. This figure, while high is still relatively small compared to the total cost of any manned Mars campaign, which is expected to total in the hundreds of billions of dollars. [76] Figure 18: MOSAIC annual cost for the timeframe Perhaps most importantly, the cost in any single year for this architecture stays below $500 million, as shown in Figure 18. This annual cost is critical when considering government budgets. Large spikes in cost from year to year are politically infeasible, with flat year-to-year costs being the ideal case. 7.2) Risk Analysis A risk analysis was performed of the MOSAIC architecture to identify the largest problems and hazards that could potentially delay the program. This risk analysis uses a scoring for both the probability of an event occurring, and the potential impact to the architecture if the event were to occur. This analysis is found in Tables below. Illinois Space Society 38