Regularized and simplified Monte Carlo Joint Probabilistic Data Association Filter for multi-target tracking in wireless sensor networks

In this paper we propose to use regularized Monte Carlo-Joint probabilistic data association filter (RMC-JPDAF) to the classical problem of multiple target tracking in a cluttered area. We have used the Monte Carlo methods in order to the fact that they have the ability to model any state-space with nonlinear and non- Gaussian models for target dynamics and measurement likelihood. To encounter with the data association problem that arises due to unlabeled measurements in the presence of clutter, we have used the joint probabilistic data association filter (JPDAF). Due to the resampling stage in the MC-JPDAF, the sample impoverishment phenomenon is unavoidable and the tracking performance will decrease. So we propose to use the Regularized resampling stage instead, to counteract this effect. Finally we have used the target dynamics model as the proposal distribution in MC-JPDAF, in order to decrease the computational cost while the performance of the tracking system is nearly maintained.