Feature-Weighted Track-to-Track Association Based on Adaptive Fuzzy C-Shell Cluster

The traditional track-to-track association (track fusion) algorithm mostly focuses on single and straight tracks, while the tracks generated by maneuvering flight, like curves, have not been researched deeply. This paper reviews current techniques of track-to-track association and improves a method, based on Adaptive Fuzzy C-Shell cluster (AFCS), which can be used among those situations where target leaves curve-like tracks. This method collects data from distributed multi-sensors network to generate track features, then uses feature-weighted AFCS algorithm to achieve track fusion. The experiment shows the proposed approach performed well under certain circumstances.