Joint Performance Optimization of Wireless Networks and Control Systems

Stanford University, Stanford CA

Investigators

Abstract

Joint Performance Optimization of Wireless Networks and Control Systems Current control technology is not well suited to operation over a packet based network, and conversely, most networks are not optimized for control applications. Moreover, critical network performance parameters that affect control system performance, such as bit rates and delay and packet loss statistics, depend on various choices made in the network, such as powers allocated to wireless transmitters, bandwidths or time-slot fractions allocated to channels, and routing methods. So in addition to the challenge of working out how a control system can be made to work in a packet-based networked environment, we also have the challenge of working out how the network should be configured, or operated, so as to improve control system performance. This proposal addresses two fundamental issues that arise. Neither solves the whole problem, but each addresses a critical issue. The first topic involves joint optimization of a controller and critical communication and network parameters such as transmit powers, packet routing, and bandwidth and time-slot allocation. To simplify this problem, only the affects of these network parameters on bit rates, hence quantization noise, is considered. The second topic considered is simultaneous routing and resource allocation for wireless data networks. This is the problem of jointly finding an optimal set of flows and communication resources that maximize the performance of the network, with respect to some objective such as optimal control system performance. Here we hope to develop algorithms to dynamically allocate communication resources (such as transmit powers, bandwidth) and simultaneously optimize routing decisions.