Event-triggered Kalman consensus filter over sensor networks

Abstract

Kalman consensus filter (KCF) has been developed for distributed state estimation over sensor networks where local estimates are exchanged with time-triggered transmission mechanism. To reduce the amount of data transfer in sensor networks, the authors propose a KCF with an event-triggered communication protocol. The triggering decision is based on the send-on-delta data transmission mechanism: each sensor transmits its local estimates to its neighbours only if the difference between the most recent transmitted estimate and the current estimate exceeds a tolerable threshold. On the basis of the event-triggered communication protocol, an optimal Kalman gain matrix is derived by minimising the mean squared errors for each sensor and a suboptimal KCF is developed for scalable considerations. By using the Lyapunov-based approach, a sufficient condition is presented for ensuring the stochastic stability of the suboptimal KCF. A numerical example is provided to verify the effectiveness of the proposed filter.