Scholarlink

<p>In modern transportation, tire wear monitoring is critical for road safety, fuel efficiency, and overall vehicle performance. Tire wear compromises traction, leading to hazardous conditions and increased maintenance costs. This study explores tire wear monitoring across five conditions—25, 50, 75, 100%, and a good-condition tire—using vibration signals captured via an accelerometer. Advanced feature extraction techniques, including statistical, histogram, and autoregressive moving average (ARMA) features, were employed, followed by feature selection using the J48 decision tree algorithm. A comparative analysis of 13 tree-based classifiers demonstrated that Random Forest paired with ARMA features achieved a classification accuracy of 100% for training, cross-validation, and testing datasets, with computation times of 0&#xa0;s for the training and testing sets, and 0.03 s for cross-validation. This robust system showcases high reliability and adaptability, significantly advancing real-world diagnostics. These findings emphasize the proactive role of automated tire wear monitoring in enhancing road safety and environmental sustainability.</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

Can tire wear be detected using vibration signals?–an experimental study

  • 期刊论文

  • 发表日期：

    2025年4月5日

查看全文
  • C. V. Prasshanth,
  • V. Sugumaran

摘要

In modern transportation, tire wear monitoring is critical for road safety, fuel efficiency, and overall vehicle performance. Tire wear compromises traction, leading to hazardous conditions and increased maintenance costs. This study explores tire wear monitoring across five conditions—25, 50, 75, 100%, and a good-condition tire—using vibration signals captured via an accelerometer. Advanced feature extraction techniques, including statistical, histogram, and autoregressive moving average (ARMA) features, were employed, followed by feature selection using the J48 decision tree algorithm. A comparative analysis of 13 tree-based classifiers demonstrated that Random Forest paired with ARMA features achieved a classification accuracy of 100% for training, cross-validation, and testing datasets, with computation times of 0 s for the training and testing sets, and 0.03 s for cross-validation. This robust system showcases high reliability and adaptability, significantly advancing real-world diagnostics. These findings emphasize the proactive role of automated tire wear monitoring in enhancing road safety and environmental sustainability.