Scholarlink

<p>Driving behaviour is the main contributor to road crashes. Increasing the road incidents on Malaysian highways raised the concern about traffic safety. The emergence of Autonomous Vehicles (AVs) has the potential to reshape traffic dynamics and safety outcomes on Malaysian roads. Yet, assessing how mixed traffic of Conventional Vehicles (CVs) and AVs influences driving behaviour during incident occurrence remains not investigated. This study assessing the incident driving behaviour impact under varying driving behaviour logics (Cautious, Normal, and Aggressive) by utilising a dynamic CoExist microsimulation model integrated with a unique optimisation method based on Genetic Algorithm (GA) and COM interface feature. First, the study analysed three years of incident data using PTV VISUM to identify incident hot spots. Secondly, the study implemented PTV VISSIM's CoExist framework to simulate three different segments of three selected highways (E2 AH2, E6 ELITE, and E23 Kerinchi Link) during peak periods. Third, a total of 23 different driving behaviours from four different driving models, which are following, car-following, lane change, and lateral movement, were calibrated to reflect a comprehensive behaviour of local conditions. Finally, an incident scenario was implemented utilising VISSIM’s event-based script to evaluate bottleneck impacts on driving behaviour. Simulation outcomes demonstrate that AVs significantly improves traffic safety compared to CVs. Specifically, AVs maintained superior operational condition awareness by dynamically expanding their forward scanning and proactively increasing their minimum look back distance. This expanded perception successfully overcomes limited situational awareness and cognitive tunnel vision exhibited by human drivers, allowing faster reactions to prevent secondary collisions.</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

Impact of Autonomous Vehicles on Driving Behaviour Incident Occurrence on Malaysian Highways: a CoExist Micro-Model and Genetic Algorithm Approach

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月7日

查看全文
  • Haitham Al-Msari,
  • Suhana Koting,
  • Mustafa Attallah,
  • Ahmed El-shafie

摘要

Driving behaviour is the main contributor to road crashes. Increasing the road incidents on Malaysian highways raised the concern about traffic safety. The emergence of Autonomous Vehicles (AVs) has the potential to reshape traffic dynamics and safety outcomes on Malaysian roads. Yet, assessing how mixed traffic of Conventional Vehicles (CVs) and AVs influences driving behaviour during incident occurrence remains not investigated. This study assessing the incident driving behaviour impact under varying driving behaviour logics (Cautious, Normal, and Aggressive) by utilising a dynamic CoExist microsimulation model integrated with a unique optimisation method based on Genetic Algorithm (GA) and COM interface feature. First, the study analysed three years of incident data using PTV VISUM to identify incident hot spots. Secondly, the study implemented PTV VISSIM's CoExist framework to simulate three different segments of three selected highways (E2 AH2, E6 ELITE, and E23 Kerinchi Link) during peak periods. Third, a total of 23 different driving behaviours from four different driving models, which are following, car-following, lane change, and lateral movement, were calibrated to reflect a comprehensive behaviour of local conditions. Finally, an incident scenario was implemented utilising VISSIM’s event-based script to evaluate bottleneck impacts on driving behaviour. Simulation outcomes demonstrate that AVs significantly improves traffic safety compared to CVs. Specifically, AVs maintained superior operational condition awareness by dynamically expanding their forward scanning and proactively increasing their minimum look back distance. This expanded perception successfully overcomes limited situational awareness and cognitive tunnel vision exhibited by human drivers, allowing faster reactions to prevent secondary collisions.