Real-time Estimation of Lane Change Risks based on the Analysis of Individual Vehicle Interactions

Abstract

Lane-changing vehicles continuously interact with adjacent vehicles by adjusting longitudinal and lateral maneuvering. Safe lane changing can be regarded as an outcome of safe vehicle interactions during lane changes. An effective safety assessment of a vehicle changing lanes is essential to developing various countermeasures to prevent crashes. This study developed a methodology for estimating the risks of changing lanes in real time using vehicle trajectory data. Both a real-time risk exposure level (RREL) and a real-time risk severity level (RRSL) were devised to quantify risk levels associated with lane changing based on an analysis of vehicle interactions. Then, fault tree analysis was adopted to integrate the two indicators, RREL and RRSL. This study demonstrated the feasibility of the proposed methodology by conducting VISSIM simulation experiments. The outcome of this study can be used to not only evaluate the effectiveness of traffic management strategies in terms of lane change safety but also support the development of various advanced driving assistance systems.