Bayesian Particle Filtering

This chapter defines Bayesian particle filtering and applies it to tracking problems that are difficult to solve by Kalman filter methods. In addition, it shows that particle filters can be applied to solve problems other than tracking. The process of computing the posterior distribution on the target’s path given the measurements received in a fixed time interval $$\left[ {0,T} \right]$$ is called fixed interval smoothing and the resulting posterior distribution is the smoothed solution. This chapter presents repeated filtering, a simple and general method for smoothing particle filters. The chapter closes with examples of applying repeated filtering to find smoothed solutions to particle filter problems that are difficult to solve by any other method.