Stability-Margin-Aware Communication Scheduling for Optimal Control of Distributed Smart Networked Systems

Abstract

The design of smart networked systems distributed over communication networks with limited bandwidth, random packet losses and stale information has to ensure closed-loop stability. Control methods based on communication efficiency are primarily of the event-based type, which send information only when the state error is above a predetermined (constant or variable) threshold value. This paper presents an alternate strategy for communication with an awareness of stability in which communication is handed out as a limited (versus threshold) control resource. The selection of agents after each sampling instant is done using an aggregate quadratic Lyapunov function ???? agents, at most, are selected consistent with a limited communication budget. The resulting stability condition is written in a ‘bounded input to state stability' (ISS) formulation, where the impact of communication imperfections is captured by a bounded residual term. A 20-agent second-order network with sparse ring-plus-two-hop topology, 5% independent packet loss, one-sample information age, bounded deterministic disturbances, and actuator saturation is studied in the proposed approach. Performance is compared to periodic communication, static error-triggered communication, and dynamic error-triggered communication. A Monte Carlo simulation with eight seeds is carried out to demonstrate that the proposed method achieves approximately a 50% reduction in transmitted signals compared to dynamic error-triggered communication and a 92.2% reduction compared to periodic communication, with comparable regulation performance to the considered baselines. The examination involves the examination of residual RMS error, post settlement RMS error and the final-to-initial Lyapunov function ratio. The stability–communication Pareto frontier is established by changing the communication budget and shows the trade-off between communication usage and regulation performance. The proposed framework can be applied to distributed secondary voltage and frequency control of inverter-based distributed energy resources in islanded microgrids where communication efficiency is one of the crucial factors to ensure a coordinated operation of the microgrid. Last, theoretical extensions that still exist to complete the stability theorem, namely an explicit bound on the marginal advantage of stability, quantization, and packet loss and hardware-in-the-loop validation under delay, remain to be identified.