Hybrid SIR joint particle filtering under limited sensor resolution

The hybrid SIR joint particle filter has been developed as an effective approximation of the exact Bayesian filter for maintaining tracks of multiple maneuvering targets from unassociated measurements. This paper further develops this approach for the situation of limited sensor resolution and two maneuvering targets. For this problem the exact Bayesian filter recursion is characterized, and is subsequently used to develop a novel version of the hybrid SIR joint particle filter. For a simple example, the effectiveness of the novel approach in handling possibly merged measurements is demonstrated through Monte Carlo simulations.