Iterative-Mapping PHD filter for extended targets tracking

A novel Iterative-Mapping algorithm is proposed to alleviate the computation intractable measurement partition problem in a Probability Hypothesis Density tracking filter for extended targets. The iterative mapping procedure is based on the assumption that the target centroids can be estimated as the conditional expectation of measurements. Using sequential Monte Carlo method, this algorithm iteratively maps the target states and measurement sources each other. It can be applied to track multiple extended targets. Simulation results demonstrate superior performance of this proposed method over traditional extended target PHD filter with partition.