State-dependent Priority Scheduling for Networked Control Systems

TL;DR

This paper proposes a scalable, state-dependent priority scheduling method for networked control systems that improves performance and stability by dynamically assigning priorities based on system states, suitable for wide-area network applications.

Contribution

It introduces a novel dynamic state-based priority scheduling policy for NCS that is scalable, efficient, and compatible with existing IP network infrastructure.

Findings

Provides performance bounds and stability analysis for the scheduling policy.

Demonstrates feasibility with a real hardware implementation.

Explores the relationship between bandwidth, transmission rate, and delay.

Abstract

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.