Scholarlink

<p>This paper implements a comparison study between three machine learning algorithms namely artificial neural network, decision tree, and random forest. This comparison is applied for the classification of hand gestures, specifically “thumbs up”, “Y-letter”, and “hand at rest”, using electromyography (EMG) data to identify motions collected from Myo armbands. The results of the three algorithms show the power of machine learning algorithms and how they can detect hand gestures from EMG signals in real-time and their accuracies with overall accuracies of 91.086% for the Artificial Neural Network (ANN), 93.28% for the Decision Tree (DT), and 95.58% for the Random Forest (RF). After the training phase, the three trained machine learning models are used to construct a brain-computer interface (BCI) as a Matlab Tool Box by using Matlab App Design. The BCI is implemented as a user-friendly application for commercial users who have no idea about machine learning and Matlab syntax. The second application is also applied using the trained machine learning models for Human–Robot interaction using GUCnoid 1.0 humanoid. The paper succeeds in achieving the comparison between machine learning algorithms, the BCI and HRI applications for commercial users, academics, and the medical and prosthetics fields.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Machine learning for EMG-based gesture recognition in brain–computer interfaces and humanoid robots

  • 期刊论文

  • 发表日期：

    2025年6月17日

查看全文
  • Amir R. Ali,
  • Mohamed W. A. Ramadan,
  • Pablo Zometa

摘要

This paper implements a comparison study between three machine learning algorithms namely artificial neural network, decision tree, and random forest. This comparison is applied for the classification of hand gestures, specifically “thumbs up”, “Y-letter”, and “hand at rest”, using electromyography (EMG) data to identify motions collected from Myo armbands. The results of the three algorithms show the power of machine learning algorithms and how they can detect hand gestures from EMG signals in real-time and their accuracies with overall accuracies of 91.086% for the Artificial Neural Network (ANN), 93.28% for the Decision Tree (DT), and 95.58% for the Random Forest (RF). After the training phase, the three trained machine learning models are used to construct a brain-computer interface (BCI) as a Matlab Tool Box by using Matlab App Design. The BCI is implemented as a user-friendly application for commercial users who have no idea about machine learning and Matlab syntax. The second application is also applied using the trained machine learning models for Human–Robot interaction using GUCnoid 1.0 humanoid. The paper succeeds in achieving the comparison between machine learning algorithms, the BCI and HRI applications for commercial users, academics, and the medical and prosthetics fields.