In this work, we focus on distributed state estimation under event-triggered measurement sampling and estimator-to-estimator communication. We design a distributed Kalman-like filter, with fully asynchronous transmissions of measurements and estimates. The estimator nodes leverage the implicit information from not receiving new sensor measurements between events, resulting in stable estimates for any transmission sequence. Moreover, we show that the performance of the centralized Kalman–Bucy filter with full measurement data can be approximated arbitrarily well with our event-triggered solution, by tuning the event thresholds and the consensus gain in the filter, while reducing communication.
Perez-Salesa et al. (Sun,) studied this question.