Low-Cost Simulation and Verification Environment for Micro-Satellites

Trans. JSASS Aerospace Tech. Japan Vol. 14, No. ists30, pp. Pf_83-Pf_88, 2016 Low-Cost Simulation and Verification Environment for Micro-Satellites By Toshinori KUWAHARA, Kazufumi FUKUDA, Nobuo SUGIMURA, Tatsuaki HASHIMOTO, Yuji SAKAMOTO and Kazuya YOSHIDA Department of Aerospace Engineering, Tohoku University, Sendai, Japan (Received July 31st, 2015) The Space Robotics Laboratory (SRL) of Tohoku University has been developing micro-satellites for years and has gained experience in their development, verification, integration, and operation. The SRL has recently started the development of model-based simulation, verification, and integration environment to realize rapid and cost-effective development of reliable micro-satellites. The conceptual design and its functionality have been verified through the real-life 50-kg-class micro-satellite projects RISING-2 and RISESAT. The developed environment can be utilized in different configurations such as full-software simulation, hardware-in-the-loop simulation, and even flight operation, depending on demands in each satellite development phase. This environment is designed to be modular and flexible. The minimum hardware configuration can be a single personal computer, which enables low-cost introduction of a satellite system simulator for a wide range of projects in a variety of project phases. It can be utilized for general micro-satellite missions and possibly even much smaller space systems in the future. Micro-satellite RISING-2 is the first satellite tested by means of this environment. Functionality of the developed simulation and verification environment was evaluated by comparing the ground simulation results and flight data obtained by RISING-2. Key Words: Micro-Satellites, Simulation, Verification, Low-Cost 1. Introduction Micro-satellites are nowadays widely utilized as platforms for space technology demonstration and scientific experiments. Tohoku University has been conducting research and development of micro-satellite technologies and have experience with several micro-satellite projects such as SPRITE-SAT (renamed as RISING-1 after the launch), RAIKO, and RISING-2. 1-4) The most recent micro-satellite being developed is RISESAT (Rapid International Scientific Experiment Satellite). RISESAT is a 50-kg-class international scientific micro-satellite, which aims to demonstrate rapid and cost-effective access to space for scientific instruments. 5) Against this background, the Space Robotics Laboratory (SRL) has recently initiated an investigation on the software-assisted development and verification environment for micro-satellites to realize the rapid and cost-effective development of reliable micro-satellite systems. 6) The aim of this research is to enhance the utilization of micro-satellites and to establish the basis of a new paradigm for the future, where cost-effective and reliable small satellites are widely utilized for both research and business purposes. This paper summarizes the configuration of the developed environment, functional verification results of ground simulation capabilities, and performance evaluation results by comparing the simulation results and real-life flight data obtained by RISING-2. 2. Simulator Architecture The main purpose of this research is to develop a comprehensive but low-cost simulation environment for the rapid development, integration, and verification of reliable micro-satellite systems. The simulation environment proposed in this paper is designed in a modular manner and is highly flexible in terms of the testing configuration setup. The system architecture and possible simulator configurations are described in the following sections. 2.1. Low-cost approach for system simulator Simulation environments are common tools used in the development of space systems. Their applications and scales differ according to the type and level of verification required, and on the available budget. It is often the case that a satellite project utilizes several different simulation tools prepared separately for individual sub-systems or functional units such as power supply systems, attitude control systems, and command and data handling systems. This type of approach can be highly cost-intensive, as the development, customization, and maintenance of these simulation tools requires considerable effort in addition to the development of the satellite itself. In contrast, the approach taken in this research combines all of these possible simulation configurations into an integrated architecture with a flexible setup configured depending on the purpose of each activity. 2.2. Simulator engine The core part of the proposed simulator is the simulator engine, which executes space environment models and the satellite model in a periodical way with a predefined time interval, as illustrated in Fig. 1. The user can implement the required environment models and satellite component models with the required accuracy. Examples of the former are the gravitational field model, magnetic field model, and atomic drag model, whereas examples of the latter are the on-board computer model, magnetometer model, and GPS model. Copyright 2016 by the Japan Society for Aeronautical and Space Sciences and ISTS. All rights reserved. Pf_83

Trans. JSASS Aerospace Tech. Japan Vol. 14, No. ists30 (2016) User Defined Environment Models User Defined Satellite Component Models Fig. 1. Simulator engine block diagram. This configuration can be regarded as the simplest configuration of full-software simulation configuration, and can be executed on a single personal computer. The merit of having this configuration is that the user can initiate the development of the on-board software without the real satellite hardware components, even at the initial stage of the project phase. 2.3. Interfacing with hardware components The communication interface between the simulator and the satellite hardware components can be handled by means of an interface front-end. Each hardware component has specific communication and data/signal interfaces. In this configuration, because of the presence of hardware components, the simulator is required to run in real-time in order to deliver time-consistent data for the hardware. For this purpose, the simulator is executed at a higher frequency than that of the real hardware component. For example, if there is a sensor that delivers data at 10 Hz, the simulator updates the internal corresponding parameter with a higher frequency, e.g., 20 Hz. This is a highly simplified implementation based on the assumption that the executing computer has sufficient computational capability. In this manner, the determination of the required capability of the workstation is left to the individual users. 2.4. Real-time emulation of the interface For some applications, the real-time behavior of the hardware components is critical in terms of periodical behavior and response delay. The approach suggested in this paper is to introduce a real-time computer as the secondary computer in order to keep the simulator engine executed on a non-real-time workstation, as described in previous sections. In this way, the developed software codes can be compatible for both simulator configurations. This is illustrated in Fig. 2. Component A Component Component B Component B Component C 20Hz Interface Front-end Hardware Component A Hardware Component B 10Hz Non-real-time Computer Non-real-time Computer Real-time Computer Semi-periodical Output Periodical Output Precise Timing Down to 1μs Timing depends on temporal computational load Fig. 3. Real-time emulation of the interface: utilization of real-time computer to achieve precise timing (above), and low-cost implementation using non-real-time workstations (below). As illustrated in Fig. 3, non-time-critical interfaces can be implemented by utilizing interface converting devices such as USB-to-Serial communications. In this case, the timing of the communication depends on the temporal computational load of the executing machine, or interfacing hardware. Depending on the requirements and the available financial budget, the user can select one of these available options. Again, the simulator engine itself can be independent of the configurations. 2.5. Hardware component stimulations An important role of the simulator is to generate stimulations for hardware components, predominantly for attitude determination and control components. Stimulators such as voltage input, optical input, and magnetic field input can be connected with the simulator to have access to space environment variables. For example, optical stimulators for star trackers can be integrated to the simulator, in order to conduct optical-path-in-the-loop simulations for the functional verification of attitude determination and control sub-system. This configuration is illustrated in Fig. 4. 2.6. Dynamic motion simulator When the verification of the dynamic response of the system is of interest, one can introduce a dynamic motion table with motion cameras for objective attitude determination and feedback to the simulator engine, as illustrated in Fig. 5. Torque commands to actuators such as reaction wheels can be sent to the real hardware components. In addition, the actual angular rates can be measured by the real satellite components such as a rate gyro. Stimulator Stimulations (Voltage, Optical, Magnetic, etc.) Hardware Component Fig. 2. Interface between simulator and hardware components. Fig. 4. Hardware component stimulator. Pf_84

T. KUWAHARA et al.: Low-Cost Simulation and Verification Environment for Micro-Satellites Direct Update of Satellite Attitude Motion Cameras Torque Command Reaction Wheel Angular Rates Rate Gyro Dynamic Motion Table Fig. 5. Dynamic motion simulator. Reaction Wheels Rate Gyro Real Dynamic Response Fig. 7. Hardware configuration of the established simulation and verification environment. 3. Hardware Configuration The established simulator can be utilized in several different configurations as follows 6) : 1) Full-software simulation environment, 2) On-board computer-in-the-loop simulation environment, 3) Static closed loop simulation environment, and 4) Dynamic closed loop simulation environment. The exemplary hardware configuration established for RISING-2 and RISESAT micro-satellites is summarized in Fig. 6, and the core hardware equipment is illustrated in Fig. 7. The hardware modules are: Computer machines for the simulator: applications are executed on either a single or multiple computer machines. Computer machines for satellite operation. Modular hardware front-end: physical interface between the simulator and real satellite hardware components. Solar simulator: generates solar array power controlled by the simulator, depending on the sun s incident angle. Electronic load: device to generate simulated power consumption of components, which is controlled by the simulator. Radio frequency communication equipment: command and telemetry interfaces between the satellite operational software and the receiver/transmitter of the satellite. Star simulator: optically stimulates star sensors according to the satellite attitude in the simulator. Dynamic motion table: enables dynamic simulation and verification of the attitude control system of the satellite. 3.1. Power control system simulator Functional verification of the power control system of a satellite is one of the most important and essential tasks of system verification. In the established configuration, the solar simulator and electronic load are utilized in addition to the real power control and battery units. The output Voltage-Current property of the solar simulator and the power consumption of the electronic load can be controlled based on the environment variables calculated by the simulator. Developers can run realistic mission scenarios and evaluate the functionalities of the power control unit computer together with the actual performance of the battery unit on the ground. 3.2. Star simulator The star simulator generates an optical stimulation as the input for the star sensor hardware as illustrated in Fig. 8. The monitor system displays star configurations in the field of view of the star sensor, according to the parameters calculated in the simulation environment, such as the satellite s attitude, presence of sun light, and shading by the earth. More than one star simulator can be used in the case where the satellite is equipped with several star sensors. Tohoku University has been conducting research on small star sensors for micro-satellites for several years. 7) In this simulation environment, star sensors take images of the monitor and process the position of the stars, and consequently the attitude of the star sensor in the inertial coordinate system. The information processed in the star sensor is sent to the satellite s attitude control unit for further processing. This environment can be utilized to evaluate the influences between camera gain and exposure time settings, the existence of white pixels, and the influence of stray light. 3.3. Dynamic motion simulator The dynamic motion table can be utilized for verification of the dynamic performance of the attitude control system of the satellite. Owing to the existence of disturbance torques such as friction and limitation in the range of movement, qualitative verification can be conducted. This system has been described by Fukuda et al. 8) Fig. 6. Exemplary configuration of the simulation environment. Fig. 8. Star sensor verification environment. Pf_85

Trans. JSASS Aerospace Tech. Japan Vol. 14, No. ists30 (2016) 4. Software Configuration The architecture of this environment is as simple and flexible as possible so that it can be utilized by a wide range of satellite projects in the future. The core software application named Satellite and Space Environment Simulator (SSES), which is written using C++, consists of two different families of classes: one is Satellite Class, and the other is Space Environment Class. The Satellite Class contains all satellite component classes, as illustrated in Fig. 9, and the Space Environment Class holds all possible space environment classes such as orbit propagation, attitude integration, geomagnetic field models, and planet models. Exemplary lists of these classes for Tohoku University s micro-satellite projects are summarized in Tables 1 and 2. Depending on the requirements and demands of each application, the models can be more detailed, replaced, duplicated, or new models can be added. As illustrated in Fig. 9, the configurations of the models, including the connections between them, are designed to be identical to the real hardware configuration. 5. Functional Verification Results The software simulated on-board computer inside the simulator can execute software codes, which are portable to real hardware equipment. In this environment, the satellite system is simulated as the sum of individual component models. These component models can be replaced with real hardware in the later development phase. In this manner, software simulated models (components) and real hardware models can be treated in the same level. The model-based development approach has the merit that the developed models can be re-utilized for other projects, as in the case where one uses the same flight hardware component for different satellite applications. 5.1. Power balancing simulation Figure 10 illustrates a comparison of the power balancing simulation results in software-in-the-loop simulation and hardware-in-the-loop simulation environments. In the former simulation, the battery model is simulated in a software environment, and in the later simulation, the real battery hardware is used. This comparison indicates that this environment can conduct sound power balancing simulations without the real battery hardware as it is intended. In this manner, one can simulate operational scenarios and evaluate power balancing characteristics, even in the accelerated simulation mode. PCU *BATPtr *SCPPtr XTX OBC1 (SCU) *PCUPtr *SASPtr *GASPtr *GPSPtr *MTQPtr X/Y/Z (OBC3) SHU *XTXPtr *LLTPtr *MMCPtr *PayloadPtrs GPS MTQ X/ Y/ Z MTQ X/ Y/ Z GAS SAS (OBC2) ACU *STTPtr 1/ 2 *FOGPtr *SESPtr *RWPtr 1/ 2/ 3/ 4 MMC RW RW STT FOG SES Table 1. List of satellite component model classes. Category Class name Contents Power Supply System Command and Data Handling Communication Attitude Determination and Control SAT_PCU SAT_BAT SAT_PANEL SAT_SCU SAT_ACU SAT_SHU SAT_URX SAT_STX SAT_GPS SAT_GAS SAT_SAS SAT_STT SAT_GYRO SAT_SES SAT_MTQ SAT_RW Power Control Unit Battery Solar Panels and Array Satellite Control Unit Attitude Control Unit Science Handling Unit URX Receiver STX Transmitter GPS Sensor Geomagnetic Aspect Sensor Sun Aspect Sensor Star Sensor Gyroscope Sun and Earth Sensor Magnetic Torquer Reaction Wheel Table 2. List of space environment classes. Category Class name Contents Basic State StateTime StateSAT Data and Time Status Satellite Basic Property Status Orbit Planet StateCLA StateECI StateECF StateGS StateGeoMag StateSun StateUmbra StateMoon StateStar Kepler Orbit Status ECL Orbit Status ECF Orbit Status Ground Station Status Geomagnetic Field Status Sun Location Status Sunshine and Umbra Status Moon Location Status Star Simulator Status Attitude StateAttitude Satellite Attitude Status Gravity and Perturbation Propagation PerturJ2 PerturSun PerturMoon PerturAeroDrag PerturSolarRadiation Integration OrbitPropagator AttitudePropagator ThermalPropagator Earth Gravity Potential (J2) Sun Gravity Potential Moon Gravity Potential Atmospheric Drag Perturbation Solar Radiation Perturbation Numerical Integrator Orbit State Propagator Attitude State Propagator Thermal State Propagator LLT PAY6 Fig. 9. Software model of a satellite inside SSES. Fig. 10. Power balance simulation: relations between power generation and consumption of a satellite. Pf_86

T. KUWAHARA et al.: Low-Cost Simulation and Verification Environment for Micro-Satellites 5.2. Attitude control simulation This section illustrates the simulation results of attitude determination and control in a software-in-the-loop simulation. The configuration of the software models and the data flow between them are illustrated in Fig. 11. The on-board attitude determination and control algorithm was executed in the software model of the attitude control unit. During this simulation, the satellite conducted a nadir pointing maneuver. The simulation result is illustrated in Fig. 12, where the angle and angular velocity errors are selected as representative values of the evaluation of the simulation results. In this manner, the functionality of the application software, which is meant to be implemented into on-board computers in the later phase, can be evaluated without having the actual hardware component of the on-board computer. The software development activity can be initiated at the very beginning of the satellite project, which enables cost-effective development, as well as an improvement in software performance and reliability. Fig. 11. Configuration of an exemplary software-in-the-loop simulation: attitude determination and control algorithms were executed inside the attitude control unit software model. 5.3. Telecommunication and data handling evaluation The telecommunication and data handling system components are integrated into the satellite in the proposed simulator configuration. This can be seen in Fig. 7. The utilized telemetry and command servers are real equipment that can be used for operation of the satellite. The command and telemetry data are transferred through radio frequency interface cables between the satellite and ground support equipment. The same setup can be used for real operations after the launch of the satellite, if required, by replacing the cables with antennas. Testing the satellite system under the same conditions as it experiences in space is very important, and can reduce the risk of unexpected failure or malfunction of the satellite system in the final ground verification phase. 6. Flight Data Analysis As a demonstration of the hardware-in-the-loop simulation, as well as operational support by means of this environment, a comparison between ground simulation and real-life flight data of RISING-2 is illustrated in this section. Figure 13 illustrates the comparison between the simulated and flight measurement data of the geomagnetic aspect sensor during a nadir acquisition maneuver. Although there is a gap in flight data due to hardware performance restrictions, Fig. 13 shows good qualitative agreement in the compared values. The simulation environment can also visualize the satellite attitude and orbital situation, as illustrated in Fig. 14, which summarizes a series of output images from the real-time visualization interface during the above-mentioned nadir acquisition maneuver. The degree of accuracy of the simulation models depends on each implementation. According to this comparison, it is revealed that the established simulation and verification environment can conduct meaningful qualitative simulations of micro-satellites and can further provide operational support for real satellite missions. This satisfies the research objectives. Fig. 12. Exemplary results of attitude control simulation: angle and angular velocity errors during a nadir pointing maneuver. Fig. 13. Comparison between ground simulation data and flight data: ground simulation uses flight data as the initial condition for propagating subsequent satellite orbit and attitude. Pf_87

Trans. JSASS Aerospace Tech. Japan Vol. 14, No. ists30 (2016) 7. Conclusion Fig. 14. Hardware-in-the-loop ground simulation results: a maneuver from lost-in-space condition to nadir pointing. This paper summarized the results of the conceptual design and functional verification of the model-based simulation, verification, and integration environment of micro-satellites developed at the Space Robotics Laboratory of Tohoku University. The different possible configurations of the environment were summarized in the beginning, followed by a description of the implementation method of model-based satellite and space environment simulators. Through the development and operation activity of micro-satellites at Tohoku University, the functionality of the environment has been evaluated. This environment will realize rapid and cost-effective development and verification schemes for achieving reliable small space systems and their safe operation. Acknowledgments This research was supported by Grant-in-Aid for Scientific Research on Innovative Areas KAKENHI: 24760658 from the Ministry of Education, Culture, Sports, Science and Technology of Japan. References Fig. 15. Image of ground surface obtained by high-precision telescope of the micro-satellite RISING-2. As can be seen in Fig. 14, real-time visualization capability of the simulator dramatically enhances understanding about the situation where the satellite is placed. Information such as the satellite attitude, illumination condition of the solar panels, pointing directions of the communication antennas, and field of views of optical sensors and scientific observation instruments can be monitored. Because the simulator can also conduct accelerated simulation if it is configured as full-software simulation configuration, forecast of these parameters and behaviors are great help for effective operational planning. A representative earth surface image taken by the micro-satellite RISING-2 is illustrated in Fig. 15. RISING-2 is the first satellite whose functionalities were verified by means of the proposed simulation and verification environment in this paper. The images taken by the high-precision telescope of the RISING-2 illustrated that it has achieved the world highest ground sampling resolution in its satellite mass category. 1) Takahashi, Y., Yoshida, K., Sakamoto, Y. and Sakanoi, T.: SPRITE-SAT: A University Small Satellite for Observation of High-altitude Luminous Events, Small Satellite Missions for Earth Observation: New Developments and Trends, Springer, (2010), pp.197-206. 2) Fukuda, K., Nakano, T., Sakamoto, Y., Kuwahara, T., Yoshida, K. and Takahashi, Y.: Attitude Control System of Micro Satellite RISING-2, Proceedings of 8th Symposium on Small Satellites for Earth Observation, Berlin, (2011), pp.206-213. 3) Kuwahara, T., Sakamoto, Y., Yoshida K., Takahashi, Y., Fukuhara, T. and Kurihara J.: Mission and System of the Earth Observation Microsatellite RISING-2, Proceedings of 8th Symposium on Small Satellites for Earth Observation, Berlin, (2011), pp.512-519. 4) Yoshida, K., Sakamoto, Y., Kuwahara, T. and Takahashi, Y.: A Series of 50kg-class Micro-satellites for Advanced Science Missions, Proceedings of 8th Symposium on Small Satellites for Earth Observation, Berlin, (2011), pp.99-106. 5) Kuwahara, T., Yoshida, K., Sakamoto, Y., Tomioka, Y. and Fukuda, K.: Satellite System Integration Based on Space Plug and Play Avionics, Proceedings of International Symposium on System Integration, Sendai, (2011), pp.896-901. 6) Tomioka, Y., Fukuda, K., Sugimura, N., Sakamoto, Y., Kuwahara, T. and Kazuya Y.: Establishment of the Ground Testing Environment for Verification and Integration of Micro-satellite, Trans. JSASS, Space Tech. Japan, 12, ists29(2014), pp.tf_33-tf_38. 7) Kuwahara, T., Battazzo, S., Tomioka, Y., Fukuda, K., Sakamoto, Y. and Yoshida, K.: System Integration of a Star sensor for the Small Earth Observation Satellite RISING-2, Trans. JSASS, Space Tech. Japan, 10, ists28(2012), pp.td_1-td_6. 8) Fukuda, K., Kuwahara, T., Tomioka Y., Sugimura, N., Sakamoto Y. and Yoshida K.: Dynamic Test Table with Spherical Air Bearing for Microsatellite, 29 th International Symposium on Space Technology and Science, 2-9 June, 2013, Nagoya, Japan. Pf_88