Collision avoidance system at urban intersections using V2X communication

Driving safely through urban intersections is quite challenging for self-driving vehicles due to complicated road geometries and the highly dynamic maneuvers of oncoming traffic, which can cause a high risk of collisions. Traditional onboard sensors like cameras, radar, and lidar give limited visibility in this environment. To overcome these limitations, this paper explores implementing a collision avoidance system at urban intersections utilizing cellular vehicle-to-everything (V2X) communication. The system leverages V2X map data to identify warning and collision zones and uses vehicle-to-vehicle (V2V) communication to estimate the maneuver types and paths of approaching vehicles. The system assesses collision probabilities by calculating the intersection points of predicted paths for the ego vehicle and oncoming traffic. Depending on the level of collision probabilities, the system generates a collision warning signal to the driver or activates emergency braking when necessary to prevent accidents. We implement the system using Simulink and the RoadRunner scenario simulation tool. To evaluate the system performance in a virtual simulation environment, we create a complex urban scene consisting of intersections with traffic lights using RoadRunner. The RoadRunner high-definition (HD) map is converted into V2X map messages in compliance with T/CSAE 53-2020 standard. The RoadRunner Scenario configures signalization of the urban intersections.