Location-based Cooperative Vehicle Collision Avoidance for Unsignalized Intersections: A Multi-sensor Integration Approach

Safety assurance of vehicles is of great importance to improve the road transportation, especially at the unsignalized intersections. To eliminate safety threat that is mostly caused by vehicle-to-vehicle collisions, technologies and methodologies of cooperative vehicle infrastructure system provide sufficient capabilities for a vehicle to share relevant information with other connected neighborhood vehicles, then identify and avoid potential collisions. In the design of collision avoidance solution, precise vehicle location is the most essential but enabling factor to support vehicle situation perception and collision avoidance decisions, which promotes location-based solutions for collision avoidance using real-time information of enhanced location from connected vehicles with wireless communication links. In this paper, a multi-sensor integration approach is presented with a GPS/DR structure where lane-level roadmap database for assistance is also integrated. A four-stage framework is designed to accomplish effective conflict resolution and safety assurance. Key issues including multi-sensor fusion with map match and the location-based collision avoidance mechanism are discussed. Simulation with field data in a real intersection environment is carried out, and the results demonstrate the feasibility and effectiveness of the proposed approach with low-cost and simple configuration.