A Two-Stage Framework for CAV Platoon Formation Transformation

Connected automated vehicle (CAV) platooning transforms intelligent transportation systems but faces challenges in formation changes, including collision avoidance and external traffic interference. We propose a two-stage framework to address these issues. The offline stage employs an A*-based cooperative maneuver algorithm that uses a matrix-based formation representation and simplified vehicle maneuver rules to convert the formation search problem into an optimal path problem, generating discrete intermediate sequences. In the online stage, we introduce a formation-priority-based distributed model predictive control (DMPC) algorithm that maintains interaction consistency among vehicles by assigning priorities based on intermediate formation characteristics. To mitigate long priority chains, our method avoids unnecessary priority comparisons and employs a responsive collision avoidance strategy. Numerical simulations, including experiments with external disturbances, validate that our approach effectively plans intermediate formations. It shows a 9–15% optimality loss compared to centralized MPC and maintains optimality consistency while meeting real-time requirements compared to prioritized DMPC.