Gyroless attitude nonlinear observer for RazakSAT satellite

Abstract

Satellite attitude determination is one of the important aspects in Attitude Determination and Control System (ADCS) of a satellite. Satellite attitude is important to be determined in a satellite to be fed back to controller in accomplishing a specific satellite mission such as Earth observation, communication, scientific research and many other missions. In commonly practice of ADCS, the angular velocity and attitude information of a spacecraft are obtained respectively from measurement of gyroscopes and attitude sensors. However, gyroscopes are generally expensive and are often prone to degradation or failure. Hence this research work is intended to study the state estimation system for gyroless spacecraft. In this work, the nonlinear mathematical model of the system is derived using combination of dynamics equation of motion and kinematics equation of motion using Euler angles attitude parameter. The observability of the derived nonlinear system is investigated using Lie derivative technique to ensure the system observability. The derived nonlinear model is also validated and verified using real in-flight telemetry attitude data of RazakSAT, the Malaysian satellite was orbiting in Near Equatorial Orbit in 2009 via extended Kalman filter (EKF), the most widely used algorithm in spacecraft practice. The non-Gaussian non-white noise in estimation system based on RazakSAT telemetry attitude data also is diagnosed and analyzed in this work using statistical techniques. Finally, the performance of state estimation during gyroless condition using particle filter (PF) algorithm is studied and compared with the EKF in terms of accuracy, computational time load and robustness aspects for efficient on-board implementation. The result shows that the gyroless system is able to provide the information of angular velocity within 0.1 deg/s accuracy, which is suitable for moderate accuracy attitude determination such as during housekeeping and detumbling mode. In terms of accuracy and robustness aspects, the PF shows its ability to provide more accurate estimation in non-Gaussian and non-white noise circumstances and more robust to measurement uncertainty. Meanwhile in terms of computational time aspect, the result shows that EKF is faster than PF algorithm. For this application, the PF is strongly suggested during contingency condition of extremely inaccurate or large uncertainty measurements such as due to unexpected failure of the existing sensor. The study of gyroless system contributes as an alternative or backup system during unavailable angular velocity data resulted from faulty sensor or deliberately designed for sensor reduction which indirectly represent cost and hardware complexity reduction.