Decomposition-Based Gradient Estimation Algorithms for Multivariate Equation-Error Autoregressive Systems Using the Multi-innovation Theory

This paper studies the parameter estimation algorithms of multivariate equation-error autoregressive systems. By using the decomposition technique, the multivariate equation-error autoregressive system is decomposed into two subsystems, and a decomposition-based generalized stochastic gradient algorithm is deduced for estimating the parameters of these two subsystems. In order to further improve the parameter accuracy, a decomposition-based multi-innovation generalized stochastic gradient algorithm is developed by means of the multi-innovation theory. The simulation results confirm that these two algorithms are effective.