Research of Dynamic Parameter Identification of LuGre Model Based on Weights Boundary Neural Network
The identification problem on the dynamic parameters in LuGre model is studied. The identification method proposed uses a neural network which considers weights boundary. The network structures and weight adjustment algorithm are given out. In order to find a set of parameters to make it approach the actual one, this algorithm uses neural network identification within the bounds of the identified parameters. Compared with the non-linear least squares (NLS) parameter identification, the relative errors of parameters which are identified by neural network based on weights boundary (WBNN) are smaller, and the precision is higher. Numerical simulation is provided to show the efficiency of the proposed method.
neural networks weight boundary LuGre model dynamic parameters identification
HUO Ai-qing RUAN Yan TANG Nan Wang Yue-long Cheng Wei-bin
Shaanxi Key Laboratory of Oil-Drilling Rigs Controlling Technique Xian Petroleum University, Xian, 710065, China
国际会议
重庆
英文
259-262
2011-01-21（万方平台首次上网日期，不代表论文的发表时间）