Adaptive control and identification of nonlinear system according to the procedure of distributed parameters

Parameter identification of block oriented nonlinear systems containing continuous nonlinearities is presented. The procedure, based on the recursive least squares identification algorithm, is presented. The entire input signal area is divided into the N sections. For each section, there exists one line in the matrix of parameters. The use of covariance matrix resetting and the combinations of forgetting factors yield good simulation results. The procedure is used for adaptive control of nonlinear processes. The results obtained are demonstrated for a model air flow control device. Advantages and disadvantages of the proposed solution are discussed.