Central Difference Particle Filter Applied to Transfer Alignment for SINS on Missiles

Abstract

For the strapdown inertial navigation system (SINS) on vertically launched and warship-borne missiles, the transfer alignment is an effective approach to estimate its navigation attitudes at the time of launching missiles, which is also called the initial navigation attitudes of SINS. The quaternions are adopted to describe attitudes, and a transfer alignment model with this description is established. However, due to the strong nonlinearity of the alignment model, the non-Gaussian distributions of gyros drifts, and the demands for alignment speed and precision, it poses a great challenge to the estimation of the initial navigation attitudes of SINS. In order to solve this problem, a new particle filter (PF) named central difference particle filter (CDPF) is introduced and applied to the transfer alignment. In this new filter, the central difference filter is used to generate proposal distribution for sequential importance sampling. A comparison study regarding the performance of CDPF with those of the extended Kalman particle filter (EKPF) and the unscented Kalman particle filter (UKPF) is conducted. The simulation results show the superiorities of the proposed approach over EKPF and UKPF.