Real-Time Control of Multiple Vehicles on a Single-Track Road with Logical Signaling

This paper presents an innovative method for autonomous driving of multiple vehicles on a single track, based on a logical vector signaling system. By integrating advanced artificial intelligence and computer vision technologies, the system achieves fully autonomous coordination of vehicle movements. The approach begins with the strategic selection of approach points along an underground track, forming connected segments. A vector-based signaling mechanism — guided by the real-time status of vehicles and a shortest-path algorithm derived from the Floyd–Warshall method — orchestrates multi-vehicle behaviors such as sequencing, reversing, and overtaking. Furthermore, the YOLO deep learning algorithm is employed for real-time object detection and avoidance, enabling dynamic route planning and smooth navigation in underground settings. This method not only ensures safe and efficient traffic flow but also significantly reduces the risk of collisions.