Over-Approximation State Estimation for Networked Timed Discrete Event Systems With Communication Delays and Losses

Abstract

This article investigates the state estimation for a networked timed discrete event system, where a plant communicates with a supervisor via a multichannel network characterized by bounded delays and losses. To address delays and losses in observation channels, we augment the plant by integrating the dynamics of these channels, thus capturing the system’s open-loop behavior. To tackle delays and losses in control channels, we augment the supervisor by considering all control decisions with potential impact on the system’s behavior. By integrating the augmented plant and supervisor, we introduce a compensated system that enables the derivation of an over-approximation of the closed-loop system’s behavior. Ultimately, we devise an online over-approximation state estimation algorithm for the closed-loop system, to compute all possible system states under communication delays and losses. We provide a simulation example to illustrate the efficacy of the proposed method.