A gyro drift and dynamics state estimator for satellite attitude control systems utilizing a star mappert†

Abstract

A recursive estimator is presented which, when given attitude measurements from a star mapper and rate integrating gyro, together with control signals, provides filtered estimates of gyro drift, disturbing angular acceleration, body angular velocity and attitude. Attention is restricted at present to satellites with substantially rigid body dynamics and negligible inter-axis coupling. A single axis attitude control system is considered, this being sufficient for demonstration of the principles. The estimator is based on the Kalman filter with the following developments ; (a) A low level fictitious covariance matrix element is introduced to facilitate estimation of the disturbing angular acceleration, and (b) the gain matrix is modified to enable the real-time model iterations and gyro sampling to be carried out at a much higher rate than the star mapper sampling. Furthermore, the part of the gain matrix relevant to the star mapper output is adapted to enable the correction to the model to be applied continuously, thereby avoiding stepping of the state estimate following new star mapper measurements. A form of the estimator is presented in which it is unnecessary to know the time to the next star mapper reading. Digital simulation results are given for a complete control system in which two control loops are completed by state-variable feedback, comprising a linear gyro-drift control law and a non-linear gas jet control law.