Multiple scatterer tracking in high range resolution radar

A target of interest measured by a high range-resolution radar sensor may be modeled by multiple dominant points of reflections referred to as scatterers. In this paper a state space model governed by static motion parameters is used to represent the motion and measurements of the scatterers moving in two dimensions. A dynamic Bayesian method, based on a particle Monte Carlo Markov chain technique known as the particle marginal Metropolis-Hastings sampler, is used to jointly infer the states and static motion parameters of the motion model. This numerical Bayesian estimation approach may be used to aid in automatic target recognition and can be used to accurately perform motion compensation in ISAR processing.