Bayesian track-before-detect algorithm with target amplitude fluctuation based on expectation–maximisation estimation

Abstract

Bayesian track-before-detect is an efficient approach to detect low observable targets. Before implementing Bayesian track-before-detect, one needs to exactly ascertain the target motion model and the measurement model. When the target return amplitude fluctuates, the target return amplitude of the measurement model is not known a priori. In the scenario, standard Bayesian track-before-detect algorithms such as particle filters, which assume perfect knowledge of the model parameters, cannot work well. In this study, the authors propose an expectation–maximisation (EM) algorithm for Bayesian track-before-detect with target amplitude fluctuation, in which the fluctuation models are incorporated into the likelihood function, and the average target return amplitude is estimated by the EM algorithm. The simulation results show that the average target return amplitude can be estimated by the EM algorithm, which is helpful in improving the performance of detection and tracking. Therefore it is feasible to apply the EM algorithm to Bayesian track-before-detect with target amplitude fluctuation.