Investigating Classification Parameters for Continuous Myoelectrically Controlled Prostheses
Investigating Classification Parameters for Continuous Myoelectrically Controlled Prostheses
复制标题
研究连续肌电控制假肢的分类参数
DOI：
--
复制
发表时间：
2005
期刊：
影响因子：
--
通讯作者：
A. Chan
中科院分区：
文献类型：
--
作者：
A. R. Goge;A. Chan
在骨骼肌收缩期间，各个肌纤维中存在相关的运动离子。这种电活动可以使用位于感兴趣的肌肉上方的表面电极来记录。产生的信号是电极附近肌纤维动作电位的总和，称为肌电信号（MES）。 MES 可用于多种应用，包括假肢控制、监测肌肉疲劳和自动语音识别 (ASR) 系统。迄今为止，MES 假肢应用中所做的研究采用了基于不同特征提取和分类算法的不同分类技术。本研究的目的是解决以下问题： 1. 假设我们使用自回归（AR）系数作为信号特征，AR 模型阶数对 MES 分类精度有什么影响？ 2. 需要多少个通道才能保持较高的分类精度？为了回答这些问题，我们进行了一系列实验来收集和处理使用不同 AR 系数作为信号特征的 MES，并使用不同的 AR 模型阶数和不同的 MES 通道集进行计算。
During the contraction of skeletal muscle, there is an associated movement ions in the individual muscle fibres. This electrical activity can be recorded using the surface electrodes located above the muscles of interest. The resultant signal is the sum of the muscle fibre action potentials in the vicinity of the electrodes, termed the myoelectric signal (MES). MESs are used in a variety of applications including prosthetic control, monitoring muscle fatigue, and automatic speech recognition (ASR) systems. The research done thus far in the prosthetic application of MESs employs different classification techniques based on different feature extraction and classification algorithms. The purpose of this research is to address the following questions: 1. Assuming we are using autoregression (AR) coefficients as signal features, what effect does the AR model order have on the MES classification accuracy? 2. How many numbers of channels are required to maintain a high degree of classification accuracy? To answer these questions, a series of experiments were performed to collect and process the MESs using different AR coefficients as signal features, computed using different AR model orders and different sets of MES channels.