Adaptive- Transition-Set Track Before Detect Algorithm Based on Dynamic Programming
Dan Le, Huang Qiang, Liu Zhaolei, Xianyan Wu · 2019
Original dynamic programming based track-before-detect (DP- TBD) assumes that targets run with nearly uniform motions and known velocities. However, targets in real-world scenes do not always run with nearly uniform motions and their velocities are usually unknown, and all we know is the minimum and maximum speeds of these targets. Original DP- TBD suffers significant performance loss under the circumstances. Aiming at the problem, this paper proposes an adaptive-transition-set DP- TBD (ATS-DP- TBD) algorithm, which refines the transition set of DP-TBD by estimating the states of targets as DP- TBD processes the measurement frames. The transition set initially is rough, which is determined by the minimum and maximum speeds, then becomes more accurate by a filter. Simulation results show that the proposed ATS-DP-TBD outperforms the original DP-TBD, the detection probability increases by 0.17, and the number of the false alarms is reduced over 94%.