State-dependent Priority Scheduling for Networked Control Systems
For designers of networked control systems, this work offers a practical scheduling approach that leverages existing IP networks, though it is incremental as it builds on known priority queuing concepts.
This paper addresses optimal scheduling for networked control systems sharing a communication channel, proposing a suboptimal policy using dynamic state-based priorities that is scalable and efficient for IP networks. It provides performance bounds and asymptotic stability, and demonstrates feasibility with a hardware implementation.
Networked control systems (NCS) have attracted considerable attention in recent years. While the stabilizability and optimal control of NCS for a given communication system has already been studied extensively, the design of the communication system for NCS has recently seen an increase in more thorough investigation. In this paper, we address an optimal scheduling problem for a set of NCS sharing a dedicated communication channel, providing performance bounds and asymptotic stability. We derive a suboptimal scheduling policy with dynamic state-based priorities calculated at the sensors, which are then used for stateless priority queuing in the network, making it both scalable and efficient to implement on routers or multi-layer switches. These properties are beneficial towards leveraging existing IP networks for control, which will be a crucial factor for the proliferation of wide-area NCS applications. By allowing for an arbitrary number of concurrent transmissions, we are able to investigate the relationship between available bandwidth, transmission rate, and delay. To demonstrate the feasibility of our approach, we provide a proof-of-concept implementation of the priority scheduler using real networking hardware.