CAJ | 학술논문

针对LuGre摩擦模型参数难以辨识的问题，提出了一种一次性获得模型所有参数的新型快速识别方法，将LuGre模型构造为以速度为输入、摩擦力为输出的单输入单输出系统，求解出相应的LuGre离散递推方程，解决了参数辨识时鬃毛微位移无法有效检测的难题。在此基础上，通过遗传算法对给定的LuGre模型进行辨识实验，发现一次性能辨识出LuGre模型的6个参数，参数辨识误差最大为0?419％，具有较高的精度。通过Simulink对指定参数和辨识参数的摩擦力仿真对比可知，辨识出的参数可以较高精度逼近原始数据，最大力矩误差仅有0?14 N·m，验证了新型辨识方法的有效性。
침대LuGre마찰모형삼수난이변식적문제，제출료일충일차성획득모형소유삼수적신형쾌속식별방법，장LuGre모형구조위이속도위수입、마찰력위수출적단수입단수출계통，구해출상응적LuGre리산체추방정，해결료삼수변식시종모미위이무법유효검측적난제。재차기출상，통과유전산법대급정적LuGre모형진행변식실험，발현일차성능변식출LuGre모형적6개삼수，삼수변식오차최대위0?419％，구유교고적정도。통과Simulink대지정삼수화변식삼수적마찰력방진대비가지，변식출적삼수가이교고정도핍근원시수거，최대력구오차부유0?14 N·m，험증료신형변식방법적유효성。
Aimed at the tough problem of identifying LuGre friction model parameters, a novel and fast parameters identification method was proposed for obtaining of the model in one?time. A single?input?output system based on LuGre model was constructed, in which speed was used as input and friction as output. Accordingly, discrete recursive equations of LuGre model were solved, and the problem of effective detecting mane micro displacement was discussed. Based on it, experiments on given LuGre model using genetic al?gorithm were conducted. For the one?time identification, six parameters within the LuGre model were identified with the maximum error of 0?419% and quite high accuracy. Through comparing the identified parameters and specified ones by SIMULINK, this system shows a high accuracy of the identified parameters approached to the original data. The maximum torque error is only 0?14 N·m, which has verified the effectiveness of this novel identification method.