Goal-Oriented Scheduling and Control Co-Design in Wireless Networked Control Systems
Jie Cao, Ernest Kurniawan, Amnart Boonkajay, Sumei Sun · 2023
We consider a wireless networked control system (WNCS) for real-time applications. Different physical processes (e.g., dynamic plant state and Markov Process-based context information) are monitored by multiple sensors via imperfect wireless links and then received/estimated by the controller for remote control. In this paper, we use violation probability as a key performance metric, which captures extreme events that violate the preset threshold of the plant state, to ensure the stability of WNCS. Scheduling multiple sensors with limited resources is challenging, due to the delays and loss in transmission, and their heterogeneous impacts on the WNCS goal. To address this problem, we first show the relationship between the considered goal and the actions of the scheduler and controller. We then present a Markov Decision Problem (MDP) to minimize violation probability with cost constraints. By analyzing the significance of information that incorporates goal-related usefulness and contextual importance, the structural results of the formulated MDP is presented. Then a goal-oriented scheduling and control co-design policy is proposed to improve the violation probability-cost tradeoff. Simulation results show that the proposed policy results in a lower violation probability and a lower cost.