Distributed Event-Driven Filtering Over Switched Sensor Networks With Reachable Set Strategy
针对拓扑切换的传感器网络中受有界非随机扰动影响的离散时间系统，提出一种基于可达集的分布式事件驱动滤波方法，将滤波误差限制在有界闭集内并优化扰动抑制水平。
This study investigates the distributed event-driven filtering problem for a class of discrete-time systems in sensor networks (SNs) with switching topology. The addressed systems are considered to suffer from the unknown nonrandom perturbations with bounded peak, and thus the existing filtering or estimation approaches for switched SNs become inapplicable. To tackle this situation, a novel reachable-set-based distributed filtering strategy is established. With the proposed piecewise Lyapunov function approach and the minimum dwell time switching mechanism, sufficient conditions are then formulated for the existence of admissible filters in both the secure and nonsecure communication channels. We prove that the resultant filtering error is ultimately confined to a bounded closed set, which can also be minimized to achieve an optimal perturbation attenuation level in the sense of reachable set. Finally, the practicability and effectiveness of the developed design technique are demonstrated via the distributed event-driven filtering of a two-spring-mass mechanical system monitored by the SN.