An adaptive technique is described which relies upon identifying the transition and driving matrices of a linear dynamic system. By identifying these matrices a sub‐optimal controller is updated so that effective control is maintained. The method may be extended to non‐linear and time varying systems by employing quasi‐linearization. The simplicity of the algorithm allows real‐time control of systems for which there is no a priori knowledge concerning system dynamics. The scheme is illustrated using a hybrid computing system specifically designed for this purpose. Application of this technique to plant is referred to. This paper gives a history of the development of this particular deterministic approach to the adaptive control problem.
CitationDownload as .RIS
MCB UP Ltd
Copyright © 1973, MCB UP Limited