Autonomous celestial navigation for lunar explorer based on genetic algorithm particle filter

Autonomous celestial navigation system is a typical nonlinear, non-Gaussian dynamic system. Extended Kalman filter (EKF) is widely used in spacecraft navigation. It only uses the first order terms in the Taylor series expansion. To nonlinear and non-Gaussian system, EKF may introduce large estimation error. Particle filter (PF) is a computer-based method for implementing a recursive Bayesian filter by Monte Carlo simulations. PF is an effective solution at dealing with nonlinear and/or non-Gaussian problems. The performance of PF relies on the choice of importance sampling density and resampling scheme. To overcome the particle degeneration and sample impoverishment problems existing in traditional particle filter method, a new autonomous celestial navigation method for lunar explorer based on genetic algorithm particle filter method is presented. Simulation results demonstrate the validity and feasibility of this new method.