Unscented Kalman Filterを用いた故障衛星のロバストな姿勢運動推定

抄録

The attitude motion estimation of objects in orbit, especially uncontrolled satellites owing to failures, is an elemental technology in space environmental preservation missions, such as debris removal and on-orbit servicing. This paper proposes a marker-based attitude estimation approach that enables the accurate estimation even when there exists uncertainty in observation and a lack of prior information such as inertial parameters. For estimating states in nonlinear systems such as attitude motion, the Unscented Kalman Filter is used because of high-accurate performance. Simulation results indicate that this method is able to achieve simultaneous estimation of states and parameters successfully in nonlinear systems, and is robust against misreading by an observer.