ROEKF-MPC Estimator for Satellite Attitude and Gyroscope Bias Estimation

Microelectromechanical system (MEMS) gyroscopes are commonly used for satellite attitude determination and control system in recent years. They are packaged in a small form factor and have lower power consumption. They provide a low-cost solution to the emerging NewSpace industry. MEMS gyroscopes exhibit time-zero null bias with variation over temperature. To overcome this problem, this article proposes a technique for self-calibration of gyroscope bias based on a reduced-order extended Kalman filter (EKF) working alongside an estimator based on the model predictive control (MPC) approach. The proposed technique is referred to as the ROEKF-MPC estimator. Both simulation results and experimental verification using an in-house developed spacecraft simulator are presented. Unlike the EKF method, the proposed method can estimate the gyroscope bias instantly and it is robust against changes in temperature. In addition, the proposed method employing a reduced-order EKF is 28.4% computationally more efficient than the typical higher order EKF method. Results show that the pointing performance is better than 0.35 degrees making the proposed method very attractive for most NewSpace applications. While the proposed method allows for self-calibration of gyroscope bias, its performance is affected by the gyroscope noise and requires the satellite's attitude to be nearly in a steady state. This limitation will be addressed in detail.