This paper investigates the state estimation problem for continuous‐time nonlinear stochastic systems in sensor networks. For this purpose, a distributed consensus filter (DCF) is proposed based on Lyapunov stability theory for every node in a sensor network. It will be proved that the mean square of the estimation error is exponentially ultimately bounded. This filter can estimate the states of nonlinear stochastic systems, the nonlinear functions of which satisfy a pseudo Lipschitz condition. Sufficient conditions for the existence of this filter are sensor network connectivity and LMI solvability of DCF. Furthermore, a criterion is presented to optimize the filter gain based on minimizing the upper consensus bound of estimation error. Simulation results show the promising performance of the proposed filter.

No abstract

No abstract