Out-of-sequence measurement algorithm based on fast Marginalized Particle Filter

When it comes to the Out-of-Sequence Measurement (OOSM) problem with nonlinear system, the particle filter (PF) is widely used. But these OOSM-PF algorithms are facing the computation burden. In order to reduce the storage and computation requirements, a new algorithm based on the fast Marginalized Particle Filter (FMPF) for the OOSM problem is proposed in this paper. By using this algorithm, the state vectors are divided into two parts: the nonlinear and linear parts. The OOSM-PF is used to deal with the nonlinear parts, while the linear parts are estimated by Kalman filter (KF) based algorithm. The algorithm solves the OOSM problem under the framework of forward directly updating. It can deal with both the 1-step-lag and the multistep lag OOSM problem. Theoretical and simulation results show the effectiveness of the algorithm in dealing with the OOSM problem.