Stability and convergence properties of model-reference adaptive control systems

Abstract

An approach to adaptive control system design which does not involve a direct measurement or identification of the variable process parameters is investigated. Convergence-time relationships and sufficient conditions for system stability are developed when the process is characterized by a variable gain or variable time-constant, and fixed dynamics of any order. The adaptive capability is achieved by employing a model as a reference element and by means of appropriate adaptive circuitry. The adaptive circuitry is simple, easily instrumented, and does not require successive differentiations; the primary function of the adaptive circuitry is one of integration.