Hybrid Adaptive Control
by
Waldemar de Castro Leite Filho
Instituto de Aeronáutica e Espaço / CTA - Brazil
Coauthors: Domingos Salvio Carrijo (Instituto de Aeronáutica e Espaço)

In continuous adaptive systems, the plant operates in continuos time and the controller parameters are adjusted continuously. Similarly, in discrete adaptive systems, the plant is modeled in discrete time and the various signals, as well as the control parameter vector, are defined at discrete instants. However, practical systems, for a variety of reasons, may contain both discrete and continuous elements. It may be desirable to make infrequent adjustments of control parameters at discrete time instants, even as continuous signals are being processed in real time, making the overall system hybrid. Even in purely discrete time systems, the adjustment of the adaptive control parameters may be carried out at rates significantly slower than the rate at which the system operates. Whether a discrete updating algorithm is used in a continuous or discrete plant, an important consideration is whether the overall system will have bounded solutions. In this paper we discuss such hybrid systems that are stable. When there are no disturbances present in the system, the convergence of the parameter vector to its true value can be arbitrarily fast by increasing the gain or the persistent excitation of the reference input. The frequency of parameters adjustments is discussed regarding the robustness of the system in the presence of disturbances.

Date received: January 26, 2000