Content area
Abstract
The problem of identification and adaptive control of retarded systems - that is those with pure time delay in cascade with a rational transfer function - is investigated in this thesis. The study is limited to single-input, single-output systems; however, the algorithms proposed can be extended to multivariable systems with time delay. In contrast to conventional discrete time approach to identification and adaptive control, this work is based on a continuous time approach. Two main algorithms are developed for identification of time delay systems. The first one is based on the polynomial identification which is a special case of nonlinear estimation technique and the second one is analogous to discrete time identification method where the degree of system transfer function's numerator is expanded to accommodate for the dead time. Both techniques involve derivative free implementation by using the concept of state variable filter. These algorithms are used in the design of self tuning control for known and unknown but constant or slowly varying time delay systems. The idea of emulator which is closely linked to predictive control is used as the basis for self tuning control. A new explicit PID adaptive control is also developed which can be used for systems with either known or unknown time delay. A new language called `CONIC' is used to develop an integrated environment for building a simulation package which is used to verify the performance of each of these algorithms.
