Event-Triggered Distributed H∞ State Estimation with Packet Dropouts through Sensor Networks
Derui Ding, Zidong Wang, Guoliang Wei · 2018
In Chapter 4 and Chapter 8, distributed filtering issues have been investigated for two classes of stochastic systems subject to network-induced phenomena or cyber-attacks. In these two topics, the data exchange is executed in a periodical paradigm, which widely neglects the behavior of communication scheduling. Recently, due to the use of scheduling, the event-triggered control and filtering issue has attracted ever-increasing research interest. Such a situation results into two challenging issues identified as follows. First, a sensor network is often subject to various network-induced phenomena (e.g., packet dropouts and stochastic nonlinearities) even if the eventtriggered communication protocol is exploited. So, we need to develop a reasonable model to describe event-triggering communication mechanisms and network-induced phenomena in a unified framework. Second, the key issue in designing distributed estimators for sensor networks is how to fuse the information available for the estimator both from itself (without event-triggered mechanism) and from its neighbors (with event-triggered mechanism). In other words, it is very critical to construct a suitable distributed estimator such that the information from different sources is adequately integrated.