Maneuvering target tracking via dynamic-programming based Track-Before-Detect algorithm

Abstract

Owing to a high detection possibility and a simple kinetic model, track-before-detect (TBD) processorsare capable to detect low signal-to-noise ratio (SNR) targets uniformly moving with constant velocities. However, when a target with weak echo is accelerating, redirecting or decelerating, conventional TBD method might be ineffective for two reasons: heavier computational cost and higher possibility of forming false trajectories. In order to solve the problem of poor capability in tracking with the maneuvering targets, we propose a dynamic programming based TBD algorithm. In this TBD procedure, higher threshold is selected in order to reduce the possibility of forming false tracks during multi-frame processing. Additionally, resulted lower detection probability can be tolerated. The performance of tracking maneuvering objects based on this TBD processor is also exhibited.