Predictive state feedback control of network control systems

Abstract

Networked control systems have gained attention in the recent years due to their widespread applications to various real time systems. Controlling these systems poses several challenges which are currently still being investigated. A study of these issues is provided along with recent proceedings in technology to counter such issues like limited bandwidth, time delays and packet drop-outs. This thesis focuses on the problem of time delays in network control system which can cause instability of closed loop operation of these systems. A guaranteed cost approach is employed to achieve stability along with achieving a certain level of performance as defined by the cost function. A state feedback controller is used and along with it, a predictive control scheme is implemented to design variable gains of the feedback controller depending on the number of packets missed (packet drop-outs) and time delays of the received input sample or state of the plant, both of which can be random but bounded for a given communication channel. The controllers are connected to the plant via the network. They generate the appropriate input for the plant so that delays in the channel will not instabilize the system and thus they comprise the network delay compensator. The controller gains and the observer gain are determined by formulating a linear matrix inequality (LMI) problem and solving this problem by using the Robust Control Toolbox in MATLAB. Further, this technique is implemented on a fictitious system by modelling the networked system with constant delay in SIMULINK and the observer states as well as the plant output are shown to be stable.