Recursive Identification Method For Two Input Nonlinear System

Nonlinear systems are widespread in the fields of chemical processes,industrial manufacturing,communication systems,and biomedicine.Its parameter identification problems have always been difficult and hot spots in the field of identification.This paper is based on recursive least squares algorithm and stochastic gradient algorithm,combined with decomposition technology,using hierarchical identification principles and auxiliary model ideas to study the identification problem of input nonlinear systems,and achieved the following research results:(1)For the special equation error system under the interference of white noise,the system is transformed into two models through the matrix transformation and the principle of hierarchical identification using the decomposition technique,combined with the principle of least squares search and negative gradient search Hierarchical least squares algorithm and hierarchical stochastic gradient algorithm for input nonlinear finite impulse response system.(2)For the special equation error system under the interference of colored noise,the interactive estimation theory in the principle of hierarchical identification is used to deal with the unmeasured noise items in the information vector,all the data information is fully utilized,and the estimation accuracy is improved.The hierarchical augmented least squares algorithm and the hierarchical augmented stochastic gradient algorithm for two-input nonlinear controlled moving average system are derived.(3)For the two input nonlinear controlled autoregressive moving average system,according to the auxiliary model identification idea,the estimated residuals are used to replace the unmeasured noise items in the information vector,and the two input nonlinear controlled autoregressive moving average is derived The system's hierarchical generalized augmented least squares algorithm and hierarchical generalized augmented stochastic gradient algorithm.