A Learning Method for System Identification

概要

論文の詳細を見る
A method for identification of a linear time invariant system is proposed which is based on the error-correcting training procedure in learning machines, and is referred to as "learning identification".1) Being applicable to cases where the input signal for the system identified is random and nonstationary, the method has the advantage of not disturbing the normal operation of the system significantly.2) Since application of the method reguires only a short time, the method can be used to identify linear quasi-time-invariant systems in which some parameters vary relatively slowly in comparison with the time reguired for identification, and noise disturbances will be eliminated by means of the moving-average method, especially if the "repeating method" of the learning identification is employed. Computer simulation of the method was carried out and periods of time required for identification were obtained for various cases. Some modifications of the method-"quantizing method" and "repeating method" of the learning identification-were also investigated together with their computer simulations, and a brief account is presented to the identification of multi-dimensional linear systems.
公益社団法人計測自動制御学会の論文