Optimal Particle Filter based on Bayesian Solution for Multiple Out-of-Sequence Measurements

In this paper, the multiple out-of-sequence measurements (OOSMs) update problem with arbitrary orders in a nonlinear and non-Gaussian system, is considered. By exploiting the complete in-sequence information (CISI) strategy, we derive an exact Bayesian solution (EBS) starting from the Bayes' rule, which provide a general solution for multiple OOSMs problem. Moreover, a particle filter implementation is developed, called CISI-EBS-PF, so as to accommodate the challenging tracking scenarios, e.g., the nonlinear and non-Gaussian system. Simulation results confirm that the tracking accuracy of the proposed algorithm is superior to the existing PF-based algorithm and quite close to that of the in-sequence processing where no delays exist in measurements.