摘要
为辨识压电驱动器的速率相关迟滞特性,提出了一种基于Takagi-Sugeno(T-S)模糊系统的迟滞建模方法。首先,引入了新的模糊C回归聚类算法(Novel fuzzy C-regression model algorithm, NFCRMA)划分模糊空间,利用模糊C均值聚类算法(Fuzzy C-means,FCM)完成初始化,再利用改进的模糊C回归聚类算法(Fuzzy C-regression model, FCRM)确定模糊前件。其次,引入了与NFCRMA相对应的超平面隶属函数,利用递推最小二乘法完成模糊后件的辨识。最后,利用上述方法对压电驱动器的迟滞特性进行动态建模。该模型能够较好逼近输入频率在10 Hz到200 Hz的迟滞特性;在正弦衰减信号作用下,模型的相对误差控制在1%之内。实验结果表明该方法有效。
A hysteresis modeling method based on Takagi-Sugeno(T-S) fuzzy system was proposed to recognize rate-related hysteresis characteristics of piezoelectric actuators.Firstly, a novel fuzzy C-regression model(Novel fuzzy C-regression model algorithm, NFCRMA) was introduced to divide the fuzzy space, and initialization was completed by the fuzzy C-means clustering algorithm(Fuzzy C-means, FCM).Then, the improved Fuzzy C-regression model(FCRM) was utilized to determine the fuzzy antecedent. Secondly,the hyperplane membership function corresponding to NFCRMA was introduced, and the recursive least squares method was used to identify the fuzzy post-component.Finally, the hysteresis characteristics of the piezoelectric actuator were dynamically modeled by the above method.The model can well approximate the hysteresis characteristic with the input frequency between 10 Hz and 200 Hz. Under the action of sinusoidal attenuation signal, the relative error of the model is less than 1%. The results demonstrate the effectiveness and practicality of the proposed modeling technique.
作者
谢慧
赵新龙
XIE Hui;ZHAO Xinlong(Faculty of Mechanical Engineering & Automation,Zhejiang Sci-Tech University,Hangzhou 310018,China)
出处
《浙江理工大学学报（自然科学版）》
2019年第5期644-649,共6页
Journal of Zhejiang Sci-Tech University(Natural Sciences)
基金
国家自然科学基金项目(61273184)
浙江省自然科学基金项目(LY15F030005)
