A model-referenced adaptive control system using a sensitivity model.

Abstract

Plant parameter variations due to environmental changes present a problem in any control system design. In this paper, two adaptive control techniques which compensate for large or small parameter variations are proposed. The variations in the parameters are identified by using the error between the plant output and a fixed model output together with the plant sensitivity functions which are used to identify plant parameter variations. One adaptive technique uses the identified parameter value to physically change the plant parameter and the other adaptive technique uses the identified parameter value to generate a compensating input to the plant. A mathematical model for the simultaneous generation of the desired output and the sensitivity functions is described. Several examples using both techniques are considered.