Improved regularized particle filter algorithm for SINS/SAR integrated navigation

This paper presents a new improved regularized particle filter algorithm for SINS/SAR (Strap-down Inertial Navigation System / Synthetic Aperture Radar) integrated navigation system. By adopting MCMC (Markov Chain Monte Carlo) move to the regularization process, a MCMC based filtering algorithm is developed through combining local resampling with MCMC move to prevent a large number of particles from being rejected. The proposed particle filtering method prevents the degeneracy of particles and guarantees that the resultant particles have a common distribution with the practical probability function, without causing extra noises on the estimates. It also reduces the estimation variance and the computational load. By using improved regularized particle filter algorithm and extended Kalman filter algorithm, simulate for the SINS/SAR integrated navigation system. Experimental results demonstrate that the improved regularized particle filter algorithm can reduce the navigation positioning errors, and filtering performance obviously exceeds the extended Kalman filter algorithm.