Gaussian Particle Swarm and Particle Filter for Nonlinear State Estimation

R.A. Krohling (Germany)


Sequential Monte Carlo Simulation, Particle Filter, Gaussian Swarm, Particle Swarm Optimization, Recursive Nonlinear State Estimation.


Sequential Monte Carlo (SMC) simulation also known as Particle Filter (PF) has successfully been applied to solve non-linear state estimation problems in the last few years. Basically, the algorithm consists of a population of “particles”, which are sampled to estimate the posterior probability distribution. The Sampling Importance and Resampling (SIR) algorithm has similarities with Gaussian Particle Swarm Optimization (GPSO). In this paper, the similarities between SIR and GPSO algorithms are identified and a concept taken from the Gaussian Particle Swarm algorithm is inserted into the PF. Preliminary simulation results are presented and compared with the standard SIR algorithm.

Important Links:

Go Back