Session
Session II: C&DH
Abstract
The CubeSat investigating Atmospheric Density Response to Extreme driving (CADRE), developed by the Michigan eXploration Laboratory, carries the Wind Ion Neutral Composition Suite (WINCS). WINCS monitors the response of Earth's upper atmosphere to auroral energy inputs and requires a specific attitude within 1 degree of pointing accuracy.
To satisfy the required pointing accuracy, multiple estimation and control algorithms are required in the attitude determination and control system (ADCS). For instance, if a satellite adopts a reaction wheel assembly (RWA) as its main control actuator, an additional sub-control algorithm and actuator set is required for desaturation of the reaction wheels. If the sub-algorithm is not managed properly, its usage could reduce the pointing accuracy established by the main control algorithm due to the conflicting objectives. Additionally, when an Extended Kalman Filter (EKF) is used for on-orbit attitude determination, a primitive estimation algorithm, such as quaternion estimation (QUEST), must be used prior to activating the EKF to reduce the initial estimation error. For such a system, the switching condition from the primitive algorithm to the EKF is critical as it can determine the performance of the filter on estimation accuracy. In order to manage the potential conflicts and switching conditions that occur in implementing multiple estimation and control algorithms, an active ADCS can benefit from adopting a hybrid system concept which is based on a finite-state machine. Dennis et al[1]. showed the effectiveness of applying a hybrid system to single and multiple satellites capable of performing orbit corrections.
To analyze the performance of the hybrid estimation and control algorithms, a hardware-in-the-loop simulation (HILS) with a 3D hemispherical air bearing was developed and tested against the results of mathematical simulations. To facilitate the HILS, and as part of the operations of CADRE, an ADCS middleware was developed based on a real-time operating system (RTOS) and timer interrupt service routine (ISR). The design of the middleware emphasizes minimization of sensor/actuator access delay to improve performance.
This paper aims to introduce and summarize our development of a multi-algorithmic hybrid ADCS system for a CubeSat, explain the procedure we used to verify this hybrid system by use of simulations and a HILS, and present the on-flight results of such a hybrid system on the CubeSat CADRE which is currently awaiting deployment onboard the International Space Station (ISS).
[1] Dennis, Louise, et al. "Satellite control using rational agent programming." Intelligent Systems, IEEE 25.3 (2010): 92-97.
Multi-algorithmic Hybrid Attitude Determination and Control System of the CubeSat "CADRE"
The CubeSat investigating Atmospheric Density Response to Extreme driving (CADRE), developed by the Michigan eXploration Laboratory, carries the Wind Ion Neutral Composition Suite (WINCS). WINCS monitors the response of Earth's upper atmosphere to auroral energy inputs and requires a specific attitude within 1 degree of pointing accuracy.
To satisfy the required pointing accuracy, multiple estimation and control algorithms are required in the attitude determination and control system (ADCS). For instance, if a satellite adopts a reaction wheel assembly (RWA) as its main control actuator, an additional sub-control algorithm and actuator set is required for desaturation of the reaction wheels. If the sub-algorithm is not managed properly, its usage could reduce the pointing accuracy established by the main control algorithm due to the conflicting objectives. Additionally, when an Extended Kalman Filter (EKF) is used for on-orbit attitude determination, a primitive estimation algorithm, such as quaternion estimation (QUEST), must be used prior to activating the EKF to reduce the initial estimation error. For such a system, the switching condition from the primitive algorithm to the EKF is critical as it can determine the performance of the filter on estimation accuracy. In order to manage the potential conflicts and switching conditions that occur in implementing multiple estimation and control algorithms, an active ADCS can benefit from adopting a hybrid system concept which is based on a finite-state machine. Dennis et al[1]. showed the effectiveness of applying a hybrid system to single and multiple satellites capable of performing orbit corrections.
To analyze the performance of the hybrid estimation and control algorithms, a hardware-in-the-loop simulation (HILS) with a 3D hemispherical air bearing was developed and tested against the results of mathematical simulations. To facilitate the HILS, and as part of the operations of CADRE, an ADCS middleware was developed based on a real-time operating system (RTOS) and timer interrupt service routine (ISR). The design of the middleware emphasizes minimization of sensor/actuator access delay to improve performance.
This paper aims to introduce and summarize our development of a multi-algorithmic hybrid ADCS system for a CubeSat, explain the procedure we used to verify this hybrid system by use of simulations and a HILS, and present the on-flight results of such a hybrid system on the CubeSat CADRE which is currently awaiting deployment onboard the International Space Station (ISS).
[1] Dennis, Louise, et al. "Satellite control using rational agent programming." Intelligent Systems, IEEE 25.3 (2010): 92-97.
