基于多源信息融合的膝上假肢步态识别方法  

Recognition method based on multi-information fusion for gait patterns of above-knee prosthesis

作　　者：高云园[1] 佘青山[1] 孟明[1] 罗志增[1] 

机构地区：[1]杭州电子科技大学自动化学院智能控制与机器人研究所,杭州310018

出　　处：《仪器仪表学报》2010年第12期2682-2688,共7页Chinese Journal of Scientific Instrument

基　　金：国家863计划(2008AA04Z212);国家自然科学基金(60903084);浙江省教育厅科研计划(Y200805157);浙江省自然科学基金(Y1090968);浙江省自然科学基金(Y1090761)资助项目

摘　　要：为了实现膝上假肢的有效控制,提出基于多源信息融合的步态识别方法。首先通过搭建人体下肢多源运动信息系统获取下肢表面肌电信号、腿部角度信号和足底压力信号。针对获取的信息,采用基于小波变换的空域相关滤波对肌电信号进行消噪并提取信号特征;选择大小腿、膝关节角度作为腿部角度信号特征;将足底压力信号通过阈值法提取有效特征。在特征提取基础上,分别利用BP神经网络和有限状态机对下肢运动信息进行步态识别,并将识别结果进行融合。实验验证了该方法在平地行走、上下楼梯模式下步态识别准确率均达到95%以上。To realize effective control of above-knee prosthesis,a recognition method based on multi-information fusion is proposed to recognize gait patterns.First,motion information acquisition system of lower limb is established to obtain EMG,leg angle and plantar pressure signals.Aiming at this information,spatial correlation filtering based on wavelet transform is adopted to eliminate noises and extract features of EMG.Thigh angle,shank angle and knee joint angle are used as the characteristics of the leg angle.Threshold method is used to process plantar pressure information and extract effective features.Based on the feature extraction,BP network and FSM are applied to recognize gait patterns respectively.Then the recognition results are fused.Experimental results show that using the proposed method,gait patterns of different motion modes,including floor walking,up and down stair climbing,are recognized effectively with recognition rate greater than 95%.

关 键 词：表面肌电信号 腿部角度 足底压力 信息融合 步态识别 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]