Chernoff Fusion of Gaussian Mixtures Based on Sigma-Point Approximation

Gunay, Melih
Orguner, Umut
The use of a sigma-point approximation is proposed for the Chernoff fusion of Gaussian mixtures. The superiority of the proposed method to its best alternatives in the literature is shown on both a simple density fusion problem and a distributed multiple target tracking example, where the outputs of Gaussian mixture cardinalized probability hypothesis density filters are fused.