Track to Track Association Based on Iterative Reference Pattern

Track to track association (TTTA) is a key problem in the multi-sensor data fusion system. Traditional TTTA algorithms are based on the absolute position information of targets degrade in the presence of sensor biases, which make the position estimates of targets deviate from their true values. Reference pattern (REP), which uses global structural feature of the spatial distribution of targets, is insensitive to sensor biases. However, false tracks can bring extraneous interferences to REP, while missed detections can result in partial loss of the structural feature. This paper, thus, presents a new TTTA approach based on iterative reference pattern (IREP). On one hand, the proposed method inherits the insensitivity of REP to sensor biases. On the other hand, it alleviates the adverse impact of false tracks and missed detections on REP by introducing the iterative process, which refines the reference track set (RDS) dynamically. Simulation results verify the performance superiority of the proposed method.