Fault Tolerant Attitude Estimation for a Nanosatellite Using Adaptive Kalman Filter with Single Scaling Factor

Abstract

In this study an integrated adaptive TRIAD/Extended Kalman Filter (EKF) attitude estimation algorithm is presented, in which the TRIAD and an adaptive EKF are combined to estimate the attitude of a nanosatellite. The quaternion set produced by the TRIAD is provided as input to the adaptive EKF. Adaptive EKF estimates the final quaternion set and using a Single Scaling Factor (SSF), it readjusts the measurement noise covariance matrix in case of a sensor fault. The performance of the presented algorithm is tested against two different fault types as noise increment and continuous bias in attitude sensors. As a result of simulations, it is seen that although the performance of the conventional EKF reduces significantly in case of sensor faults, adaptive EKF continues to give reliable attitude estimations.

Keywords

• nanosatellite, attitude estimation, adaptive kalman filtering, fault, magnetometer, sun sensor