Particle Filtering for Attitude Estimation Using a Minimal Local-Error Representation

Abstract

A computationally efficient spacecraft attitude estimation particle filter is derived. The quaternion is used to represent the global attitude and the modified Rodrigues parameters are used to represent the local errors. The mean attitude is computed through the weighted sum of the modified Rodrigues parameter particles. The idea of progressive correction is employed to ensure the proper operation of the particle filter. Simulation results indicate that the performance of the particle filter exceeds that of the extended Kalman filter or the unscented Kalman filter for very large initialization errors.