A modified particle filter for nonlinear systems with application to tracking problem

This paper presents a modified recursive Bayesian estimation algorithm that combines an importance sampling based measurement update step with a bank of Sigma-Point Kalman Filters for the time-update and proposal distribution generation. The posterior state density is represented by a Gaussian mixture model that is recovered from the weighted particle set of the measurement update step by means of a weighted Expectation-Maximization (EM) algorithm. This step replaces the resampling stage needed by most particle filters and mitigates the sample depletion problem. A tracking example shows that this new approach has a better estimation performance than standard particle filter.