Collision-Free Multi-vehicle Trajectory Planning with Vehicle Meeting Constraints on Continuous Road

In autonomous transportation systems such as unmanned mines, connected and autonomous mining trucks are prone to accidents when meeting in high-risk zones such as narrow curves. To improve safety in autonomous traffic systems, this research introduces a novel approach for multi-vehicle trajectory planning with a specific focus on avoiding vehicle meeting within high-risk zones situated along a continuous road. We mainly focus on the short-term trajectory planning for road segments with one high-risk zone, and the problem is mathematically formalized as a mixed integer linear programming (MILP) problem. The objective function is set as a weighted sum of total delay and uncomfortableness, and the problem can be efficiently solved by the Gurobi optimizer within one second. By leveraging the previously planned trajectory as the input for the subsequent trajectory planning problem for the next high-risk zone, the method can facilitate continuous trajectories for vehicle across the entire road with multiple hazardous zones (e.g., on mountain roads). Numerical experiments conducted under variable traffic demand scenarios illustrate the effectiveness of the proposed method in averting vehicle meetings within high-risk zones while maintaining a notably low total delay.