改进的自适应神经模糊推理系统的角度传感器误差补偿方法  

Improved adaptive neural-network-based fuzzy inference system angular sensor error compensation method

作　　者：王艳永[1,2,3] 邓方[1,3] 孙健[1,3] 

机构地区：[1]北京理工大学自动化学院,北京100081 [2]中国北方车辆研究所,北京100072 [3]复杂系统智能控制与决策教育部重点实验室,北京100081

出　　处：《控制理论与应用》2013年第10期1342-1346,共5页Control Theory & Applications

基　　金：省部级重点基金资助项目(9140A17051010BQ0104);北京市教育委员会科技成果转化与产业化资助项目(20110639023)

摘　　要：角度传感器测量精度控制在工程应用中非常重要,直接影响其实际应用的效果.当被测物理量和角度传感器输出之间为复杂非线性关系时,传统方法已难以获得满意的结果.本文引入了一种基于改进的自适应神经模糊推理系统的误差补偿方法,阐述了模型建立过程与步骤,并对一个16位绝对式光电编码器进行了精度检测与误差补偿.实验结果证明,与多项式拟合法和BP神经网络相比,改进的自适应神经模糊推理系统可显著提高光电编码器的测量精度;相比于补偿前,补偿后光电编码器测量精度可至少提高7.5倍.The control of the measurement accuracy of angular sensors is very important in engineering applications; and has significant effects on the operation performances of the applications. Traditional methods cannot provide satisfactory results when the input-output relationship of angular sensors is complex and nonlinear. To deal with this problem, we propose the error compensation method based on the improved adaptive neural-network-based fuzzy inference system （ANFIS）. The modeling procedures are demonstrated step by step in this paper. This method has been applied to calibrate a 16-bit absolute-type photoelectric encoder based on the accuracy test. The results show that, compared with the polynomial fitting and BP neural network, the improved ANFIS enhances the measurement accuracy markedly. The measurement accuracy of the optical encoder is raised up to at least 7.5 times higher than that of the original value.

关 键 词：角度传感器 自适应神经模糊推理系统 误差补偿 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]