Algorithm for Solving Volterra Integral Equations Met in Process Identification

摘要

This report is concerned with the development of an effective solution to the problem of the determination of the impulse response of a physical system as applied to the problem of identification of process dynamics in adaptive control systems. With input command signals whose characteristics are changing with time, an optimal control of the process can be achieved most effectively by altering the control characteristics at appropriate instants during the length of the operation. A general approach to the problem is chosen so that complete identification is obtained using only the normal input signals without relying on the injection of test signals. An effective procedure has been developed to process these measured data to obtain the entire impulse response. The proposed procedure, with some modifications and generalization of concepts, can be applied to the problem of designing the impulse response of an optimum linear system using minimum mean squared error criterion. (Author)