Maximum Likelihood-based Multi-innovation Stochastic Gradient Method for Multivariable Systems

Abstract

This paper considers the parameter estimation problems for multivariable controlled autoregressive moving average systems. By means of the decomposition technique, a multivariable system is transformed into several identification submodels according to the number of outputs. A maximum likelihood extended stochastic gradient identification algorithm is derived for identifying each subsystem by using the maximum likelihood principle. In order to improve the convergence rate, a multivariable maximum likelihood-based muti-innovation stochastic gradient algorithm is proposed. The proposed algorithms can generate more accurate parameter estimates compared with the multivariable extended stochastic gradient algorithm. The illustrative simulation results show that the proposed methods work well.