一种基于T-S模糊辨识的人体站立平衡系统建模方法  

A Modeling Method for Human Standing Balance System Based on T-S Fuzzy Identification

作　　者：王洪瑞[1,2] 刘琨[1] 肖金壮[2] 熊鹏[1] 

机构地区：[1]燕山大学电气工程学院,秦皇岛066004 [2]河北大学电子信息工程学院,保定071000

出　　处：《生物医学工程学杂志》2014年第6期1243-1249,共7页Journal of Biomedical Engineering

基　　金：国家自然科学基金资助项目(61074175)

摘　　要：为了使被动平衡康复训练系统可以制定安全的训练强度和训练方式,本论文采用T-S模糊辨识方法建立了一种新的人体站立平衡系统数学模型。该模型以多维运动平台的运动加速度为模型输入,人体关节角度为模型输出。采用人工蜂群寻优算法改进模糊C-均值聚类算法,提高了辨识前件参数的效率。通过试验,采集了9位健康成年人的被动站立平衡调节数据,用于模型参数训练和模型结果验证。采用仿真结果和测量数据的均方差及互相关度,证明了所建模型是准确和合理的。In order to develop safe training intensity and training methods for the passive balance rehabilitation train- ing system, we propose in this paper a mathematical model for human standing balance adjustment based on T-S fuzzy identification method. This model takes the acceleration of a multidimensional motion platform as its inputs, and human joint angles as its outputs. We used the artificial bee colony optimization algorithm to improve fuzzy C- means clustering algorithm, which enhanced the efficiency of the identification for antecedent parameters. Through some experiments, the data of 9 testees were collected, which were used for model training and model results valida- tion. With the mean square error and cross-correlation between the simulation data and measured data, we concluded that the model was accurate and reasonable.

关 键 词：被动站立平衡 T-S模糊模型辨识 模糊C-均值聚类 数学模型 

分 类 号：R318[医药卫生—生物医学工程]