Scholarlink

<p>The development of automotive intelligence has become an unstoppable global trend. To keep up with the forefront of human–machine interaction design and meet the growing demand for driving comfort from consumers, this study comprehensively considers multiple factors such as the driver's body shape, muscle activation, and driving behavior to establish four comfort prediction models. Through comparison, it was found that the support vector machine (SVM) leads with an <i>R</i><sup>2</sup> of 92.18%, having the lowest RMSE and Theil IC, making it the overall optimal model. The RBF model slightly outperforms in MAPE, but is generally inferior to SVM. The decision tree model ranks the lowest across all metrics. Based on the prediction models, the driver's body pressure distribution data is obtained through a pressure mat, and the driver's muscle activation levels are assessed using AnyBody simulation. Finally, an orthogonal experimental design is used to determine the optimal size parameters for the cockpit layout, including the seat, pedals, and steering wheel. </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

Optimising Driver Comfort Through Pressure Distribution and Physiological Data: A Predictive Model for Human–Machine Interface Design

  • 期刊论文

  • 发表日期：

    2025年6月5日

查看全文
  • Mingyue Li,
  • Fei Yu,
  • Baohua Ding,
  • Chongyue Shi,
  • Qi Wang,
  • Yisheng Du,
  • Hong Li

摘要

The development of automotive intelligence has become an unstoppable global trend. To keep up with the forefront of human–machine interaction design and meet the growing demand for driving comfort from consumers, this study comprehensively considers multiple factors such as the driver's body shape, muscle activation, and driving behavior to establish four comfort prediction models. Through comparison, it was found that the support vector machine (SVM) leads with an R2 of 92.18%, having the lowest RMSE and Theil IC, making it the overall optimal model. The RBF model slightly outperforms in MAPE, but is generally inferior to SVM. The decision tree model ranks the lowest across all metrics. Based on the prediction models, the driver's body pressure distribution data is obtained through a pressure mat, and the driver's muscle activation levels are assessed using AnyBody simulation. Finally, an orthogonal experimental design is used to determine the optimal size parameters for the cockpit layout, including the seat, pedals, and steering wheel.