A multi-model track-before-detect algorithm for manoeuvring target detection for Over-the-Horizon Radar

In this paper a multi-model track-before-detect (MM-TBD) is proposed for the HF Over-The-Horizon-Radar (OTHR) manoeuvring target detection. Multiple hypotheses about target manoeuvres and echo sources are represented in the multi-model-estimation framework with each sub-filter corresponding to one hypothesis. The track output of each sub-filter is obtained based on the maximum likelihood criterion and then a generalized likelihood ratio for hypothesis test is presented to target detection. A numerical simulation is presented to show the effectiveness of the proposed algorithm.