Coupled Iterative Identification Methods For Multivariable Output-Error-Like Systems

With the rapid development of modern industrial technology,more and more variables and parameters are included in the control system in the industrial process,so the parameter identification of this multi-variable control system has become a hot research in recent years.Multi-variable output error system is a special kind of multi-variable system.With its complicated parameter types,it becomes a research hotspot in the field of system identification.This paper conducts research on multivariable output-error-like systems,and their parameter identification problems under different colored noise disturbances.Using the coupling identification concept,a series of coupling identification algorithms have been studied.The deduced algorithms have theoretical significance and practical application value.The paper has achieved the following results:1.For the multivariable output-error-like moving average systems,the extended gradient-based iterative algorithm is derived based on the auxiliary model identification method.Furthermore,in order to improve the accuracy of the parameter estimation,the partially-coupled extended gradient-based iterative algorithm is studied by making use of the coupling identification concept.2.For the multivariable output-error-like moving average systems,the extended least squares-based iterative algorithm is derived based on the least squares search principle.In order to improve the performance of the algorithm,the coupling identification concept is used as an optimization strategy,and the partially-coupled subsystem extended least squares-based iterative algorithm is proposed.3.For the multivariable output-error-like autoregressive moving average systems,the generalized extended gradient-based iterative algorithm and the generalized extended least squares-based iterative algorithm are studied.In order to improve the calculation efficiency,the corresponding subsystem generalized extended iterative algorithms are studied based on the hierarchical identification principle.4.For the multivariable output-error-like autoregressive moving average systems,the partially-coupled generalized extended gradient-based iterative algorithm and the partially-coupled subsystem generalized extended least squares-based iterative algorithm are proposed.These two algorithms effectively improve the accuracy of parameter estimation and the convergence speed by enhancing the coupling between the same parameter estimates between subsystems.In summary,this paper studies the coupled iterative identification algorithms for multivariable output-error-like systems with two different noise disturbances.The MatLab platform is used to analyze the above algorithms,each of the above algorithms are simulated and analyzed,and the effectiveness of the algorithms is proved.Then,the performance of the coupled iterative identification algorithm under different noise variances,and different data lengths are studied by adjusting the simulation conditions.The effectiveness of the algorithms is verified.In the final summary and outlook section,a brief introduction is given to some of the research priorities and difficulties in this direction,and some generalizations of the algorithm studied in the paper are also given.