HRA*-ORCA: A Novel Framework for Initial Guess Generation in Cooperative Lane Changing of CAVs

Efficient trajectory optimization is crucial for connected and automated vehicle (CAV) systems, especially in cooperative lane-changing scenarios. This paper addresses the limitations of traditional methods—static, non-cooperative solutions from A* and the lack of global optimality in local collision avoidance approaches like Optimal Reciprocal Collision Avoidance (ORCA). We propose a multi-stage framework, HRA*-ORCA, which integrates the global planning capabilities of Hazard Reactive A* (HRA*), Minimum Jerk refinement for smooth trajectories, and ORCA for resolving inter-vehicle conflicts. The resulting initial guesses are collision-free, dynamically feasible, and cooperative. Simulation results in multi-vehicle highway scenarios highlight significant improvements in both optimization efficiency and trajectory quality compared to traditional methods, providing a promising solution for optimization-based CAV cooperative lane-changing trajectory planning.