A Novel Track-Before-Detect Method Based on H-PMHT and the Polynomial Filter

In order to improve the detection performance of radar for maneuvering weak targets in complex battlefield environments, a histogram probabilistic multi-hypothesis tracking (H-PMHT) method based on the polynomial filter is proposed in this paper. This method uses histogram model to directly synthesize target measurement from multi-frame radar observation data, so as to advance the signal-to-noise ratio. Considering the maneuverability of weak targets, we combine the enhanced memory polynomial (EMP) with characteristic of self-initialization and the faded memory polynomial (FMP) which can track indefinitely as the pre-filter, use the Kalman filter as the main filter (EMP/FMP-KF) to obtain nonlinear state estimation from the mixed measurement and mixed proportion generated by the H-PMHT algorithm. Simulation results show that the EMP/FMP-KF based H-PMHT algorithm is an effective track-before-detect method for maneuvering weak targets, and it has greater advantages in tracking accuracy than the Kalman filter based H-PMHT algorithm.