Simulation and Automation of Lateral Collision Scenarios Using Traffic Simulation

Lateral collisions whether during lane changes or straight-line driving are a major safety concern on roads. These crashes often happen due to blind spots, sudden lane shifts, or even simple drifting, leading to dangerous and sometimes fatal accidents. This study uses the Simulation of Urban MObility (SUMO) software to model high-risk scenarios where vehicles collide sideways, whether during intentional lane changes or unexpected deviations while driving straight. By simulating interactions between an autonomous Ego vehicle and human-driven cars, this study analyzes what makes these lateral collisions so common and how they might be prevented. To capture the subtle movements that lead to sideswipe crashes, SUMO’s Sub Lane model is used, which allows for precise tracking of how vehicles shift sideways not just during lane changes but also when they drift out of position. Analysis of different high-risk situations, such as aggressive lane changes, poor gap judgments, and sudden swerves are done to see how often they result in near-misses or actual collisions. Trajectory data from Floating Car Data (FCD) helps validate these simulations, making the findings more reliable. Since manually testing thousands of traffic scenarios would be impractical, automation is implemented using Python, running countless simulations with different traffic parameters and driver reaction times. This process provides massive amounts of data on how, when, and why lateral and longitudinal collisions occur. The insights from this research can help make roads safer, whether by improving autonomous vehicle controls, refining driver assistance systems, or guiding better traffic policies. By understanding the root causes of these crashes, we can work towards reducing them in both human-driven and self-driving traffic.