ABSTRACT The partially‐coupled information vector system refers to a multivariable system in which some of the information (inputs) among sub‐systems are coupled, while the rest of the information (outputs) are uncoupled (independent). After parameterizing such a multivariable system, we obtain a partially‐coupled information vector identification model. Based on this coupled identification model, we propose corresponding iterative parameter identification methods, including hierarchical gradient‐based iterative algorithms, hierarchical least‐squares‐based iterative algorithms, hierarchical multi‐innovation gradient‐based iterative algorithms, and hierarchical multi‐innovation least‐squares‐based iterative algorithms, etc. These iterative parameter identification methods can be extended to other linear and nonlinear multivariate stochastic systems with colored noise.
Ding et al. (Wed,) studied this question.