A New Approach to Cost-Reference Particle Filtering

In this paper we propose a new cost-reference particle filter that exploits the concept of correlative learning. The objective of applying correlative learning is to obtain the centers of probability distributions that are used for particle generation. Such distributions should provide particles in regions of the state space that have low costs. The new cost-reference particle filter is compared to the original one through computer simulations of a target tracking system that uses range and bearings-only sensors, which are colocated.