In this work, a mixed Kalman/H-infinity filter is designed for the attitude estimation of a low Earth orbit microsatellite and theexternal disturbance torques acting on it. The state vector will be formed by satellite...In this work, a mixed Kalman/H-infinity filter is designed for the attitude estimation of a low Earth orbit microsatellite and theexternal disturbance torques acting on it. The state vector will be formed by satellite’s attitude along with angular rates andthe external disturbances. An improved external disturbance modeled as a random walk acting (slowly varying) around threeaxis attitude was proposed. This external disturbance is mainly generated by the aerodynamic torque, the residual magneticmoment and the gravity gradient torque. The satellite has only magnetometer on board as the attitude sensor. The proposedalgorithm is tested using simulated data for a microsatellite, and the results of this study are tested in different scenarios. Thefirst two scenarios are the cases with and without uncertainty in the satellite’s inertia. The last scenario is extensive MonteCarlo simulations with uniformly distributed initial conditions of the Euler angle and angular rate. The major purpose ofthis work is to demonstrate that we can estimate external disturbances and attitude dynamic parameters of a satellite usinga simple filter that combines the best features of Kalman and H∞ filters. The simulation results show that the attitude RMSerror is less than ±1 deg (acceptable accuracy). Also, Monte Carlo simulation gives good results of the proposed filter. Thislatter estimates the attitude with accuracy less than 0.8 deg, the rate order is 1 milli-deg/s and the external disturbances around1.5 μNm.展开更多
文摘In this work, a mixed Kalman/H-infinity filter is designed for the attitude estimation of a low Earth orbit microsatellite and theexternal disturbance torques acting on it. The state vector will be formed by satellite’s attitude along with angular rates andthe external disturbances. An improved external disturbance modeled as a random walk acting (slowly varying) around threeaxis attitude was proposed. This external disturbance is mainly generated by the aerodynamic torque, the residual magneticmoment and the gravity gradient torque. The satellite has only magnetometer on board as the attitude sensor. The proposedalgorithm is tested using simulated data for a microsatellite, and the results of this study are tested in different scenarios. Thefirst two scenarios are the cases with and without uncertainty in the satellite’s inertia. The last scenario is extensive MonteCarlo simulations with uniformly distributed initial conditions of the Euler angle and angular rate. The major purpose ofthis work is to demonstrate that we can estimate external disturbances and attitude dynamic parameters of a satellite usinga simple filter that combines the best features of Kalman and H∞ filters. The simulation results show that the attitude RMSerror is less than ±1 deg (acceptable accuracy). Also, Monte Carlo simulation gives good results of the proposed filter. Thislatter estimates the attitude with accuracy less than 0.8 deg, the rate order is 1 milli-deg/s and the external disturbances around1.5 μNm.
