Scholarlink

<p>Pavement surface characteristics play a pivotal role in enhancing driving comfort and road safety, while surface texture is believed to correlate with these diverse characteristics. In this study, pavement surface characteristics, texture and relevant variables from 679 asphalt pavement sections across 46 U.S. states were extracted from the Long-Term Pavement Performance (LTPP) database. Five texture indicators and twelve explanatory variables, including pavement roughness, traffic volume, climate condition, aggregate properties mixed with gradation and volumetric properties were compiled and summarized. Comprehensive descriptive statistics were presented to understand the dataset. Afterward, the correlational heat map of these variables was generated to exclude the parameters with strong correlations and remove their potential of multi-collinearity. Statistical regression analysis and machine learning models—random forest and gradient boosting algorithm—were developed to determine significant factors of pavement texture. The most significant variables are International Roughness Index (IRI), and Annual Average Daily Traffic (AADT), followed by several other key factors including freezing index, freeze thaw cycle, and percent passing on #4 and #200 sieves, aggregate geo-classification, binder content, and percent air voids in mixes. This study can help to better understand pavement texture properties and assist in pavement condition assessment.</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

Evaluating Pavement Surface Texture with LTPP Database

  • 期刊论文

  • 发表日期：

    2025年12月11日

查看全文
  • Masud Rana Munna,
  • Kaustav Chatterjee,
  • Joshua Q. Li

摘要

Pavement surface characteristics play a pivotal role in enhancing driving comfort and road safety, while surface texture is believed to correlate with these diverse characteristics. In this study, pavement surface characteristics, texture and relevant variables from 679 asphalt pavement sections across 46 U.S. states were extracted from the Long-Term Pavement Performance (LTPP) database. Five texture indicators and twelve explanatory variables, including pavement roughness, traffic volume, climate condition, aggregate properties mixed with gradation and volumetric properties were compiled and summarized. Comprehensive descriptive statistics were presented to understand the dataset. Afterward, the correlational heat map of these variables was generated to exclude the parameters with strong correlations and remove their potential of multi-collinearity. Statistical regression analysis and machine learning models—random forest and gradient boosting algorithm—were developed to determine significant factors of pavement texture. The most significant variables are International Roughness Index (IRI), and Annual Average Daily Traffic (AADT), followed by several other key factors including freezing index, freeze thaw cycle, and percent passing on #4 and #200 sieves, aggregate geo-classification, binder content, and percent air voids in mixes. This study can help to better understand pavement texture properties and assist in pavement condition assessment.