抄録
This paper presents a design method for a model reference adaptive control system which can be applied to an unknown plant with zeros.
In this method, the plant input is generated by an adaptive controller with variable parameters so that the plant output may closely approximate the desired response given by a reference model. In this case, the ord er of the reference model can be selected to be lower than the order n of the plant by at most the number m of the plant zeros. The number of variable parameters needed for generating the plant input is n+m+1 in total. A parameter adjusting rule of “integral+proportional+differential” type is derived, based on Popov's hyperstability theorem. In this rule, the integral term is required for guaranteeing the asymptotic stability of the adaptive system. The proportional and the differential terms are effective for improving the response characteristics of the system. But in practical application the differential term may be eliminated, because its effect is not as remarkable as that of the proportional term.
This proposed method is applied to a sample system composed of a 3rd order plant and a 2nd order reference model, and the behavior of the adaptive system is investigated by computer simulation studies.
