摘要:This paper presents five distributed receding horizon filters for linear discrete-time systems with uncertainties. To design robust fusion algorithms against model uncertainties, receding horizon strategy is adopted, and five fusion algorithms are introduced: optimal fusion, convex combination, covariance intersection, median fusion, and information fusion. Optimal fusion, convex combination, and covariance intersection are based on weighted sums of local receding horizon Kalman estimates; however, information fusion is computed based on the information form of the Kalman filter which is based on weighted sums of local measurements, and median fusion comes from selection of the intermediate value among local estimates. Performance comparison in terms of accuracy is discussed through an example.