Convex Combination and Covariance Intersection Algorithms in Distributed Fusion

In a distributed estimation or tracking system, local estimates are first generated from individual sensors. The state estimates of associated objects are then fused to generate the global estimates. The fusion algorithm has to deal with correlated estimation errors due to common past information or common process noise. Most approaches to estimation fusion use a convex combination of the local estimates to minimize the mean square error. Recently an alternative approach to fusion, covariance intersection, has been proposed and is claimed to be more robust. This paper provides a set-theoretic interpretation of the covariance intersection approach and develops a tighter bound for the estimation error. Numerical results are used to compare the performance of the different fusion approaches, and the probability for the bound to be true is computed for some examples. Critical parameters that affect the performance of a fusion algorithm are identified.