Filter designing with finite packet losses and its application for stochastic systems

Packet losses are a general problem in networked control system and wireless sensor networks (WSNs). In order to increase estimation accuracy and reliability, a filter with finite packet losses called minimum variance filter (MVF) is designed for stochastic systems. The authors develop a packet loss model, and design a novel MVF based on the orthogonal analysis approach. The proposed filters rely only on the packet arrival probability at each time instant and do not need to know whether the measurement is received at a particular time instant. However, the proposed MVF is not applicable to non-linear cases. So an extended MVF is derived for stochastic non-linear systems and applied to track a moving target in WSNs. Simulation results show that, compared to existing methods, the proposed MVFs have superior performance.