Track fusion with incomplete covariance information

This work solves a practical decentralized track fusion problem where the global fusion center can not receive the real-time covariance information from local systems because of the limited communication bandwidth. The algorithms of track-to-track fusion were revisited, especially the recent new researching results, and the proper algorithms for the linear systems with incomplete covariance information were analyzed. The approximate optimal track fusion (AOTF) algorithm was proposed for steady or "slowly varying" system based on the optimal decentralized track fusion algorithm (OTF). This new designed algorithm has preferable performance than the maximum-likelihood-based fusion methods either in tracking error or in computational load, at the same time, the communication bandwidth needed is very small. Simulation results confirm its effectiveness.