Rao–Blackwellised particle filter based track-before-detect algorithm

Abstract

A track-before-detect (TBD) algorithm based on the Rao–Blackwellised particle filter (RBPF) is proposed for over-the-horizon radar (OTHR) target detection and tracking. Since the unthresholded measurements are integrated over time in the TBD algorithm, it can potentially detect and track targets with a much lower signal-to-noise ratio than conventional methods. In order to take advantage of the TBD algorithm, both system dynamic and measurement models are defined for the application under consideration. Furthermore, by exploiting the linear sub-structure available in the models, an RBPF based TBD algorithm is derived in detail. The proposed algorithm can provide reduced variance for the estimates with linear sub-structure and a lower computational complexity than the standard particle filter. Simulation results show that the proposed algorithm is capable of detecting and tracking dim targets for OTHR efficiently.