Distributed Fusion with GM-LCC Filter Based on Covariance Intersection and Weighted Association

Track-to-track fusion plays an important role in distributed fusion system, which provides accurate target state estimation. In the clutter environment, Gaussian mixture will produce many Gaussian components with little physical meaning, which may result in interruption tracks and false alarm tracks after fusion. This paper proposes an improved fusion method for distributed fusion. The Gaussian mixture covariance intersection is used to update the step in the Gaussian mixture linear-complexity cumulant filter. In the target estimation part, the weighted association is used to associate the new track points with the historical track points. Finally, the target state is estimated, and there is no interruption tracks and false alarm tracks. The effectiveness and superiority of the method are verified by simulation in the environment of dense clutter and close target distribution.