摘要
假手动作模式识别是对假手进行有效控制的基础。本文对国内外假手动作模式识别的主要方法(神经网络、模糊理论、聚类分析、支持向量机(SVM))的原理及其研究状况进行了系统的综述,特别是针对各种方法不同动作模式识别的准确率和识别结果的稳定性进行了较详细的对比分析与研究。研究结果表明,基于支持向量机(SVM)的假手动作模式识别方法在动作识别的正确率和稳定性方面具有最优的特性。
Artifical hand's movements are identified effectively based on Movement's Pattern Recognition methods for Prosthetic Hand.In this paper,the theories and research condition of Movement's Pattern Recognition methods (neural networks,fuzzy theory,clustering analysis,support vector machine (SVM)) for Prosthetic Hand are reviewed systematically.In particular,the accuracy of various pattern recognition methods and the stability of the identified results are analyzed and studied in detail.The studied result shows that the method based on support vector machine (SVM) on the accuracy and the stability of Movement's Pattern Recognition is best.
出处
《微计算机信息》
2010年第19期183-185,107,共4页
Control & Automation
基金
基金申请人:胡加华
项目名称:基于肌电信号实时比例控制的电子假手研究
基金颁发部门:上海市教委(JWCXSL0902)
关键词
假手
肌电信号
模式识别
artifical hand
electromyography
pattern recognition