Scholarlink

<p>The escalating frequency of accidents along the highway poses significant challenges, including human casualties, property damage, and traffic disruptions. Therefore, to address these concerns, the study focuses on accidents on a 53.7&#xa0;kM stretch of a national-level highway in a developing country, using data physically collected from 2021 to 2023. Initially, crash attributes associated with both fatal and nonfatal incidents have been collected to conduct a comprehensive analysis of accident characteristics. Furthermore, the study investigates “Blackspot identification” and the “Accident severity index with ranking” to pinpoint high-risk locations and assess the seriousness of accidents along this stretch. Statistical analysis, such as correlation tests, ANOVA tests, and regression analysis, of the attributes has been explored. Conversely, sensitive crash attributes have also been examined based on various vehicle types, aiding in implementing targeted safety measures and enhancing road safety. In addition, the study employs Artificial Neural Network techniques to develop crash prediction models. Through subsequent analyses, several vital attributes have been identified as significant determinants of crash outcomes (fatal/nonfatal). Notably, attributes such as the “Cause of crash”, “Road geometry”, “Month of crash occurrence”, “Time of crash occurrence”, and “Vehicle type” have emerged as particularly influential. Therefore, by delineating potential risk attributes associated with crashes, this research facilitates informed decision-making for mitigation strategies to avoid accidents, which is particularly pertinent in developing countries.</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

A Safety Assessment and Accident Prediction Over National Level Highway Using Artificial Neural Network

  • 期刊论文

  • 发表日期：

    2025年4月10日

查看全文
  • Bittu Ghosh,
  • Somnath Karmakar

摘要

The escalating frequency of accidents along the highway poses significant challenges, including human casualties, property damage, and traffic disruptions. Therefore, to address these concerns, the study focuses on accidents on a 53.7 kM stretch of a national-level highway in a developing country, using data physically collected from 2021 to 2023. Initially, crash attributes associated with both fatal and nonfatal incidents have been collected to conduct a comprehensive analysis of accident characteristics. Furthermore, the study investigates “Blackspot identification” and the “Accident severity index with ranking” to pinpoint high-risk locations and assess the seriousness of accidents along this stretch. Statistical analysis, such as correlation tests, ANOVA tests, and regression analysis, of the attributes has been explored. Conversely, sensitive crash attributes have also been examined based on various vehicle types, aiding in implementing targeted safety measures and enhancing road safety. In addition, the study employs Artificial Neural Network techniques to develop crash prediction models. Through subsequent analyses, several vital attributes have been identified as significant determinants of crash outcomes (fatal/nonfatal). Notably, attributes such as the “Cause of crash”, “Road geometry”, “Month of crash occurrence”, “Time of crash occurrence”, and “Vehicle type” have emerged as particularly influential. Therefore, by delineating potential risk attributes associated with crashes, this research facilitates informed decision-making for mitigation strategies to avoid accidents, which is particularly pertinent in developing countries.