A Dynamic Programming Track-Before-Detect Algorithm Based on EKF for Acceleration Targets

Aiming at nonlinear issue of real system model to some extent in projects and poor performance as a result of constant transfer step size in conventional Dynamic Programming Track-Before-Detect (DP-TBD) algorithm when tracking acceleration targets, a DP-TBD algorithm based on extended Kalman filter (EKF-DPTBD) is proposed in this paper. The new algorithm makes use of EKF to estimate different transfer step sizes between consecutive states, avoiding mismatch between the transfer step size and target velocity when detecting acceleration maneuvering targets and thus the state searching efficiency for acceleration targets is improved. The proposed method can track targets with both strong and weak maneuverability accurately. Simulation results verify the effectiveness of the proposed algorithm and performance of this algorithm is analyzed and compared with conventional DP-TBD algorithm.