Least Squares Based Iterative Identification For Two-Input Multirate Systems

Based on the National Nature Science Foundation of China, this thesis studies least squares based iterative identification methods for two-input multirate systems. It not only is significant in theory, but also has potentially values in applications. On the basis of the relevant literature, the history of system identification is briefly recalled and the relevant parameter estimation methods are summaried. The results are as follows:1. The two-input multirate systems are the main research object of the thesis, the state space models are derived for the multirate systems with two different input sampling periods and further the corresponding transfer functions are obtained, and different two-input multirate stochastic systems models are given with different noise models.2. The least squares based iterative algorithm of two-input multirate output error models is proposed for two-input multirate output error systems. The main idea is using auxiliary model identification techniques to replace unknown noise-free outputs in the information vector with the outputs of the auxiliary model, and replacing the unmeasurable varibles in the information vector with their iterative estimates based on the hierarchical principle and iterative method.3. For output error moving average models for two-input multirate systems, the corresponding identification models contain not only unknown variables, but also the unknown noise terms. To solve the difficulties of the identification, the least squares based iterative algorithm is presented for such multirate systems. Compared with output error moving average models, the controlled autoregressive moving average models contain more parameters obviously. The least squares based iterative algorithm for the controlled autoregressive moving average models is studied using the iterative method.4. The least squares based iterative algorithm is extended to the systems with autoregressive noise models, that is two-input multirate systems with autoregressive noises including the output error autoregressive model and controlled autoregressive autoregressive model, and the least squares based iterative algorithms are presented and the detailed calculation steps are given.In summary, the thesis mainly studies least squares iterative identification methods for two-input multirate systems. Comparded with existing recursive identification, the proposed algorithms have high parameter estimate accuracy. Finally, a simple conclusion of this thesis is given. The difficulties and future topics based on the iterative identification are simply outlined.