Network Robustness Depth and Topology Management of Networked Dynamic Systems

Networked control systems are subject to adversary conditions that affect their network topologies.To ensure reliable system operations,network topologies need to be characterized and managed for their impact on the overall system performance.This paper introduces the concept of network robustness depth for this pursuit.Discrete event systems are used as a foundation to model dynamic behavior of network topologies,support their analysis,and carry out their management.Stochastic analysis relates the link reliability probabilities to a probabilistic characterization of network robustness depth.Several topology management strategies are discussed,including passive methods,random strategies,and optimization methodologies.Their respective benefits and limitations are quantified.By using platoon control as a platform of hybrid(continuous and discrete event) systems and packet erasure channels as a communication protocol,the results are demonstrated with case studies.