Track fusion via track filter in dense sensor environments

• A novel track fusion algorithm is proposed in dense sensor environments.
• Track filter is used to select the appropriate sensor tracks for fusion.
• An optimal number of sensor tracks are fused.
• The superiority of the algorithm is verified in theory and simulation.

At present, most track fusion approaches fuse all available sensor tracks without regard of their different contributions to system tracks. However, in dense sensor environments, some abnormal sensors or conflicted information will increase the complexity and the time of track fusion greatly, even lead to opposite results. Therefore, a track fusion algorithm via track filter (TFTF) is proposed in this paper. The track filter is based on a two-stage paradigm of abnormal sensors detection and heuristic state estimation fusion. At first, the abnormal sensors are detected by Chauvenet criterion. Then the state estimation fusion is guided by the heuristic function, in which the appropriate sensor tracks are selected and an optimal number of tracks are fused. The effectiveness and superiority of the algorithm are verified in theory and simulation.