Strategies for Optimal Control in Networked Systems under Limited Communication

Abstract

The optimal control of networked control systems (NCS) with limited 
communication channels poses significant challenges, such as delays, packet losses, and 
constrained data rates, which degrade system performance. Recent advancements propose 
strategies to address the delimitations by focusing on communication-control codesign, adaptive 
scheduling, and robust optimization. A deep reinforcement learning (DRL) algorithm 
demonstrates superior scheduling and control input performance for large-scale wireless NCS, 
efficiently handling timecorrelated channels (Pangetal.,2024). For connected vehicles, range
limited and time-delay communications are addressed through non-model-based control 
methods, ensuring system stability despite un- certainties(Wangetal.,2023). The presence of 
Markovian packet losses introduces complexity in the control problem, where stochastic 
difference equations are utilized to establish necessary conditions for stability and optimality 
(Han et al., 2023; Wang et al., 2023). Additionally, trade-offs in state estimation, balancing data 
rates and observability under dropout constraints, highlight the importance of tailored strategies 
for effective NCS performance (Liuetal.,2023). Despite the advances, computational complexity 
and real-time adaptability remain key areas for further research, particularly for dynamic 
environments where stringent response times are critical.