In order to guide and control traffic flows from all directions at intersections for achieving optimal ecological operation, based on the concept of swarm intelligence cooperation, the control of traffic flows at intelligent connected intersections is transformed into a heterogeneous multi-agent swarm control problem composed of connected vehicles and traffic signals. By integrating macroscopic traffic flow characteristics of intersections with the microscopic ecological benefits of vehicles, an ecological guidance and cooperative control method for intersection traffic flow is constructed, which combines vehicle guidance with signal coordination optimization. Through a traffic flow-queue cooperative control mechanism and an iterative feedback strategy, the method generates a combination of vehicle trajectories and signal timing schemes that maximize the overall ecological benefits of the system. Furthermore, a fast solution method based on multi-agent reinforcement learning is designed to improve both the accuracy and timeliness of the control scheme optimization process.Experimental results demonstrate that the proposed model can dynamically generate vehicle guidance schemes and signal cooperative control schemes at intersections under intelligent connected environments.

connected vehicles, heterogeneous clusters, collaborative control, ecological guidance