Recursive Distributed Filter Design for 2-D Systems Over Sensor Networks: On Component-Based, Node-Wise and Dynamic Event-Triggered Scheme

In this paper, the recursive distributed filtering problem is investigated for a class of discrete shift-varying two-dimensional systems over sensor networks. To alleviate the resource consumption, a new component-based, node-wise yet dynamic event-triggered scheme is proposed to regulate data transmissions among the neighboring sensor nodes over the sensor-to-filter channels. The aim is to devise a distributed filter to ensure the existence of an upper bound on the filtering error variance and subsequently minimize such a bound at each iteration. In virtue of stochastic analysis techniques and mathematical induction principle, a recursive algorithm is developed to calculate the desired upper bound which is then locally minimized by appropriately parameterizing the filter gains. Furthermore, the effect of the event-triggered scheme on the estimation accuracy is rigorously evaluated. The validity of the established filter design strategy is finally demonstrated by a numerical simulation.