DRIVING BEHAVIOUR ASSESSMENT AND SAFETY ANALYSIS ON EXPRESSWAYS IN INDIA

Abstract

Road safety is a critical global concern with significant opportunities to save lives. Research indicates that human factors are the primary cause of road crashes. Previous studies suggest that over 90% of accidents are caused due to human errors. Given the variability and complexity of driving behaviour, there is a pressing need for innovative approaches to collect real-time exposure data encompassing driver, vehicle, road, traffic, and environmental factors. Such data are crucial for predicting crash events and assessing risk. This study focused on analyzing driving behaviour on expressways (access controlled high speed corridors) in India, characterized by inconsistent lane usage and mixed traffic conditions, to develop effective crash countermeasures. In this regard, this study set out to achieve four key objectives aimed at enhancing road safety on expressways. First, it sought to analyze risky driving behaviours using the Driver Behaviour Questionnaire (DBQ). Second, it aimed to develop a methodology for extracting surrounding environmental data using a single, cost-effective Light Detection And Ranging (LiDAR) sensor. Third, the study assessed lane change behaviour. Finally, it evaluated kinematic thresholds for various safety surrogate measures such as Time to Collision (TTC), Modified Time to Collision (MTTC), and Deceleration Rate Required to Avoid Collision (DRAC), to identify safety-critical events. To analyze driving behaviour, this study implemented a dual-method data collection strategy: (i) DBQ and (ii) the instrumented vehicle approach. Utilizing the DBQ, 546 responses were gathered through an online survey to assess driving behaviours on expressways, providing preliminary insights into the driving patterns of Indian drivers. Further, empirical data was obtained from realworld driving conditions using an instrumented vehicle equipped with advanced sensing technologies, including LiDAR, pedal force sensors, GPS, cameras, and a Video-VBOX system. This setup involved a diverse group of 60 drivers varying in age, gender, driving experience, and professional backgrounds. The resultant dataset, termed the 'Expressway Drive: Instrumented Vehicle (EDIV) Dataset', comprised approximately 8000 km of driving data in both day and night conditions, specifically curated for expressway conditions. The description of each objective is presented in the following paragraphs.