Scholarlink

<p>Detecting fatigue in the tibialis anterior muscle is essential in preventing falls among elderly individuals with muscle weakness. However, current classifiers for muscle fatigue detection face two key challenges. The first is optimizing model parameters, as improper parameter settings can significantly reduce detection accuracy. The second challenge involves identifying a reliable feature suitable for real-time detection. This study proposes a hybrid optimization algorithm, HPSOGWO-SVM, to detect current muscle fatigue status. The algorithm combines particle swarm optimization (PSO) and the grey wolf optimizer (GWO) to optimize the parameters of a Gaussian radial basis function (RBF)-based support vector machine (SVM). This approach efficiently identified the optimal parameter combination (<i>p</i> and <i>γ</i>), significantly reducing the manual search time compared to previous methods. Results from the ablation study showed that HPSOGWO-SVM consistently outperformed other combinations and traditional models in recognition performance while enhancing computational speed. The selection of surface electromyography (sEMG) features played a crucial role in the model's performance. This study extracted and calculated nine time-domain and frequency-domain features from sEMG to assess recognition accuracy. The findings demonstrated that the proposed sequential median frequency (sMDF) feature, combined with HPSOGWO-SVM, achieved a high accuracy rate of 82.23% and demonstrated real-time detection capability compared to other features.</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

Hybrid PSO-GWO algorithm with SVM integration for detecting tibialis anterior muscle fatigue

  • 期刊论文

  • 发表日期：

    2025年6月16日

查看全文
  • Fu-Sheng Lin,
  • Jyun-Rong Zhuang

摘要

Detecting fatigue in the tibialis anterior muscle is essential in preventing falls among elderly individuals with muscle weakness. However, current classifiers for muscle fatigue detection face two key challenges. The first is optimizing model parameters, as improper parameter settings can significantly reduce detection accuracy. The second challenge involves identifying a reliable feature suitable for real-time detection. This study proposes a hybrid optimization algorithm, HPSOGWO-SVM, to detect current muscle fatigue status. The algorithm combines particle swarm optimization (PSO) and the grey wolf optimizer (GWO) to optimize the parameters of a Gaussian radial basis function (RBF)-based support vector machine (SVM). This approach efficiently identified the optimal parameter combination (p and γ), significantly reducing the manual search time compared to previous methods. Results from the ablation study showed that HPSOGWO-SVM consistently outperformed other combinations and traditional models in recognition performance while enhancing computational speed. The selection of surface electromyography (sEMG) features played a crucial role in the model's performance. This study extracted and calculated nine time-domain and frequency-domain features from sEMG to assess recognition accuracy. The findings demonstrated that the proposed sequential median frequency (sMDF) feature, combined with HPSOGWO-SVM, achieved a high accuracy rate of 82.23% and demonstrated real-time detection capability compared to other features.