Paper
Confidence Region Test in Closed Loop System Identification
- Authors:
- Jian-hong Wang; Yong-hong Zhu; Xiao-yong Guo
- Abstract
- Here, we construct a new confidence region to test the quality of closed loop system identification. This confidence region with respect to model parameters is derived based on an asymptotic normal distribution of the parameter estimator and its covariance matrix, which are estimated from sampled data. The uncertainty bound of the model parameter is constructed in the probabilistic sense by using the inner product form of the asymptotic covariance matrix. Further, the statistical analysis result is used to design the optimal input signal. Thus, our result in this short paper can extend the breadth of the system identification field. Finally, the simulation example results confirm the theoretical identification results.
- Keywords
- Closed Loop Identification; Model Uncertainty; Confidence Region Test
- StartPage
- 112
- EndPage
- 120
- Doi
- 10.18005/JCSE0401011