Target tracking using multiple auxiliary particle filtering

Particle filters are a widely used tool to perform Bayesian filtering under nonlinear dynamic and measurement models or non-Gaussian distributions. However, the performance of particle filters plummets when dealing with high-dimensional state spaces. In this paper, we propose a method that makes use of multiple particle filtering to circumvent this difficulty. Multiple particle filters partition the state space and run an individual particle filter for every component. Each particle filter shares information with the rest of the filters to account for the influence of the complete state in the observations collected by sensors. The method considered in this paper uses auxiliary filtering within the MPF framework, outperforming previous algorithms in the literature. The performance of the considered algorithm is tested in a multiple target tracking scenario, with fixed and known number of targets, using a sensor network with a nonlinear measurement model.