On suboptimal Kalman filtering in case of cluttered observations

This paper examines Kalman filtering and three suboptimal methods in the case of delayed and out-of-sequence measurements in networked or wireless systems. The motivation for suboptimality for Kalman filtering stems from the low computational resources of wireless devices. Optimality is reduced by using a history of a finite horizon. It is shown that the suboptimal filtering with full iteration within the update window performs very similar to the optimal filter, even with a short history window.