Scholarlink

<p>The real-time control of prosthetic limbs using EEG signals has been recognized as a complex and underexplored challenge in biomedical engineering. To address this, a robust task recognition model was developed to improve the accuracy and responsiveness of artificial limb control. EEG data from the publicly available MILimbEEG dataset were utilized, and statistical features—Arithmetic Mean (AM), Standard Deviation (SD), and Skewness (S)—were extracted from time-domain signals across various motor tasks. These features were then refined through the ReliefF feature selection technique. A stacking ensemble model was constructed, employing Gradient Boosting and AdaBoost as base learners. Through evaluation, the model achieved accuracy, precision, and recall rates of 96.1%, 96.2%, and 96.1%, respectively, surpassing the performance of the individual Gradient Boosting (50.9%) and AdaBoost (83.3%) models. Based on these results, the proposed approach is considered suitable for integration into real-time prosthetic control systems, offering improved interpretability and responsiveness for assistive neurotechnology applications.</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

Enhancing a task recognition model for real-time control of artificial limbs using data-driven boosting-based stacking on the MILimbEEG dataset

  • 期刊论文

  • 发表日期：

    2026年1月29日

查看全文
  • Wissam H. Alawee,
  • Luttfi A. Al-Haddad,
  • Ali Basem

摘要

The real-time control of prosthetic limbs using EEG signals has been recognized as a complex and underexplored challenge in biomedical engineering. To address this, a robust task recognition model was developed to improve the accuracy and responsiveness of artificial limb control. EEG data from the publicly available MILimbEEG dataset were utilized, and statistical features—Arithmetic Mean (AM), Standard Deviation (SD), and Skewness (S)—were extracted from time-domain signals across various motor tasks. These features were then refined through the ReliefF feature selection technique. A stacking ensemble model was constructed, employing Gradient Boosting and AdaBoost as base learners. Through evaluation, the model achieved accuracy, precision, and recall rates of 96.1%, 96.2%, and 96.1%, respectively, surpassing the performance of the individual Gradient Boosting (50.9%) and AdaBoost (83.3%) models. Based on these results, the proposed approach is considered suitable for integration into real-time prosthetic control systems, offering improved interpretability and responsiveness for assistive neurotechnology applications.