Investigative Analysis of Suboptimal Filter for State Estimation in Object Tracking Wireless Sensor Network

The research presented here is the investigation of particle filter (PF) performance in a state estimation of a point mass type noncooperative moving target in an object tracking wireless sensor network (OTWSN). Motion dynamics of the target is modeled as a stochastic state-space model by simulating different time-varying parameters of a jump Markov linear system in a finite way considering these parameters are sufficiently large to meet the practical needs. Performance is evaluated on the basis of the lower bound of the posterior Cramér-Rao bound as root-mean-square error (RMSE). It has been found that the performance is subject to dynamics of the target motion and sensor observation process under consideration. The present research focuses on multiple motion dynamics, keeping in mind the real-time noncooperative moving target tracking in OTWSN. Results indicate a positive correlation between RMSE and motion models. Correlation increases as the motion dynamics change, but still, PF demonstrates suboptimal response to motion dynamics.