Session
Frank J. Redd Student Competition
Location
Salt Palace Convention Center, Salt Lake City, UT
Abstract
CubeSat missions increasingly require sustained science-grade precision pointing over long durations. Although sub-arcsecond line-of-sight (LoS) stability is achievable, maintaining this performance is challenged by nonlinear, speed-dependent, and time-varying reaction wheel behavior. As actuator degradation progresses from beginning-of-life (BOL) to end-of-life (EOL), model-plant mismatch can lead to significant pointing degradation and potential loss of mission capability. This work investigates the limitations of fixed-gain and parametric adaptive control under degraded actuator conditions and proposes a periodic identification and controller redesign framework. The approach detects persistent model mismatch using innovation-based consistency and performance metrics, triggers on-orbit identification, and updates the actuator model using a nonparametric speed-dependent lookup table integrated into feedforward compensation. The framework is evaluated in a closed-loop AOCS simulator including spacecraft dynamics, sensors, and degraded reaction wheels. For localized speed-dependent degradation, feedback-only control yields pointing errors of 100 to 300 arcsec, while adaptive feedforward fails outside the anomaly region. The proposed method reduces the relative pointing error by more than an order of magnitude, from several hundred arcseconds to within ±10 arcsec, restoring near-nominal performance. These results show that when degradation induces structural model mismatch, periodic identification and model updating are necessary to maintain precision pointing. The framework provides a practical strategy to extend mission lifetime in CubeSat-class systems.
Document Type
Event
Beyond Parametric Adaptation: Periodic Redesign for Reaction Wheel Degradation in Precision Pointing CubeSats
Salt Palace Convention Center, Salt Lake City, UT
CubeSat missions increasingly require sustained science-grade precision pointing over long durations. Although sub-arcsecond line-of-sight (LoS) stability is achievable, maintaining this performance is challenged by nonlinear, speed-dependent, and time-varying reaction wheel behavior. As actuator degradation progresses from beginning-of-life (BOL) to end-of-life (EOL), model-plant mismatch can lead to significant pointing degradation and potential loss of mission capability. This work investigates the limitations of fixed-gain and parametric adaptive control under degraded actuator conditions and proposes a periodic identification and controller redesign framework. The approach detects persistent model mismatch using innovation-based consistency and performance metrics, triggers on-orbit identification, and updates the actuator model using a nonparametric speed-dependent lookup table integrated into feedforward compensation. The framework is evaluated in a closed-loop AOCS simulator including spacecraft dynamics, sensors, and degraded reaction wheels. For localized speed-dependent degradation, feedback-only control yields pointing errors of 100 to 300 arcsec, while adaptive feedforward fails outside the anomaly region. The proposed method reduces the relative pointing error by more than an order of magnitude, from several hundred arcseconds to within ±10 arcsec, restoring near-nominal performance. These results show that when degradation induces structural model mismatch, periodic identification and model updating are necessary to maintain precision pointing. The framework provides a practical strategy to extend mission lifetime in CubeSat-class systems.
