Optimal input design for identifying parameters and orders of MIMO systems with initial values

Along with the development of modern industrial systems, the identification issues for multi-input multi-output (MIMO) systems have attracted much attention. Initial values of systems have an influence both on the input and output which directly reduces the accuracy of estimated model parameters. To solve the identification problem for MIMO systems with initial values, this paper proposes a simple and efficient identification method containing the optimal input design. The multi-dimensional inverse M-sequences are used as the input signals, which both basically meet the requirements of optimal input design and have a reference value in the practical applications. Based on the optimal input design, we can eliminate the influence of initial values generated by the system and develop an identification algorithm for multi-input, multi-output stochastic systems by means of finite impulse response (FIR) models. Finally, the identification procedure is to discuss the Pade approximation method to produce the parameter estimates, and further to use the loss functions to determine the orders of MIMO systems. The simulation results show that the proposed method is quite efficient.