A Fast Nonlinear Model Identification Method for on Line Test

摘要：

The identification of nonlinear dynamic systems using linear-in-the-parameters models for on line test is studied in this paper.A recursive algorithm to predict the estimated correction based on the principal of the least square estimation that has the structure of an optimal minium-variance filter is developed.An estimate of the measured variable is obtained in real time as weighted average of the calibrated measurements for on line test.Instead of the using the recursive least squares algorithm,an Extended Kalmn Filter(EKF) algorithm is used.EKF has the advantage of computational simplicity and is well suited to represent time varying features in real-time easurements.Simulation examples are given which show the proposed methos is numerically more stable than the approach existed.

关键词： Nonlinear dynaic system linear parameters model least squares Extended Kalman Filter recursive algorithm numerical sability.

作者: Xiuxia Du Pingkang Li Qingjun Lu

作者单位: Beijing Jiaotong University, Beijing, 100044,China Northern Vehicle Research Institute, Beijing, 10072, China

会议类型: 国际会议

会议名称: 第七届国际测试技术研讨会

会议地点: 北京

会议语种:英文

在线出版日期: 2007-08-05（万方平台首次上网日期，不代表论文的发表时间）