Design of Multi-lane Formation Transformation Method Driven by Finite State Machine

This paper studies the scenario of multi-lane vehicle formation maneuvering from any initial formation to any final formation. Firstly, a two-layer control framework is designed. The upper layer is based on the Hungarian algorithm to complete the target allocation of formation members, and the lower layer is based on the distributed model predictive control to complete the lane change maneuver of formation members. Secondly, the corresponding control barrier functions are designed for different stages of lane change as the anti-collision constraints between formation members, and the switching of constraints in different states is completed based on the finite state machine. Finally, the effectiveness of the proposed method is verified by numerical simulation.