Performance investigation on constraint sufficient statistics distributed particle filter

The constraint sufficient statistics distributed particle filter is a novel and effective solution for bearings-only single target tracking. The algorithm achieves a significant reduction in communication overhead by factorizing the likelihood function without suffering a major decrease in accuracy. However, the algorithm has some limitations which we discuss and explore in this paper. In particular, the algorithm has a bias induced via the approximate likelihood calculation, depending on the geometry of the sensors relative to the target.