Distributed sampling rate adaptation for networked control systems

Building networked control systems is a promising direction that promotes the evolution of the traditional control systems. The ability of using different sampling rates in control systems provides the flexibility for adapting their resource needs based on the dynamic networking environment. This paper studies the dynamic rate adaptation problem for networked control systems. In particular, we define a utility function which quantifies the relationship between the performance of a control system and its sampling rate. Then we formulate the rate adaptation problem as an optimal resource allocation problem, where the aggregated utility is maximized. We further present a price-based algorithm, where prices are generated to reflect the network utilization penalties and are used as the basis for rate adaptation. We formally prove the stability of our algorithm. The rate adaptation algorithm is further evaluated in an integrated simulation environment that consists of Matlab and ns-2, which allows highly accurate evaluation of network effects on the NCS performance. The experiment results show that our algorithm is able to provide agile and stable sampling rate adaptation.