摘要

This paper presents a control framework to address a typical vehicle-to-vehicle (V2V) encountering scenario of lane exchanging with two vehicles traveling on contiguous lanes in the same direction and are maneuvered to exchange lanes. A trajectory generating model for lane-changing maneuver is integrated into the driver-vehicle system. Based on this system, both the risk of collision and the lane-exchanging intentions of the drivers can be considered in the trajectory replanning. A linear time-varying model predictive control with parameters of a discretized nonlinear system is proposed to realize the optimization objectives of satisfying the drivers' maneuvering intentions and reducing the risk of collision. The optimization problem is subject to the constraints consisting of the human driver preferences, driving habits, and driving capabilities. Simulations have verified the controller's performance and robustness in different scenarios, including different initial conditions for the risk of collision and different combinations of drivers' preferences and driving habits. The results show that the proposed controller can satisfy drivers' intentions while avoiding collisions in all these scenarios.