Novel event‐triggered distributed state estimation algorithm for nonlinear systems over wireless sensor networks

Abstract

This article focuses on the event-triggered distributed state estimation for nonlinear dynamic systems over wireless sensor networks, where whether measurement should be transmitted from the sensor to the corresponding local estimator depends on a predesigned event-triggered mechanism. To obtain a better estimation performance while saving the communication energy consumption, a novel event-triggered nonlinear state estimator is designed by approximating the true posterior probability density function with minimum Kullback-Leibler divergence when the measurement is not transmitted. After local estimation results are produced in every local estimator through applying the cubature rule, according to a communication protocol and covariance intersection fusion strategy, an event-triggered distributed cubature Kalman filtering (EDCKF) algorithm is developed. Compared with algorithms based on weighted average consensus, the proposed algorithm eliminates the disagreement between local estimators and obtains its best performance within a limited time. Moreover, sufficient conditions are obtained to prove the stability of the EDCKF algorithm. Simulation results are provided to demonstrate the effectiveness and superiority of the proposed estimator and algorithm.

Event‐triggered differential private distributed fusion estimation for sensor networks with unknown cross‐correlations

“…These ET mechanisms have been verified to be much more energy efficient than time‐triggered ones 20,21 . Consequently, they have been adopted in ET control 22–26 and estimation 13,27–30 . Thus, the first motivation of this article is reducing communication consumptions of differential private estimation methods through introducing ET mechanisms.…”

“…20,21 Consequently, they have been adopted in ET control [22][23][24][25][26] and estimation. 13,[27][28][29][30] Thus, the first motivation of this article is reducing communication consumptions of differential private estimation methods through introducing ET mechanisms.…”