Critical Trajectory Point Planning for Connected and Autonomous Vehicles on Freeway On-Ramps Under Mixed Traffic Environment

With the development of the cooperative vehicle and infrastructure systems (CVIS), trajectory planning for Connected Autonomous Vehicles (CAVs) provides a promising solution to the management of merging vehicles at a freeway on-ramp. Existing studies mainly focus on the planning of trajectory profiles or trajectory points at each time step assuming CAVs can exactly follow such planned trajectories. However, challenges arise due to high computational burden and limited communication bandwidth in the near future where low-level CAVs and Regular Vehicles (RVs) coexist. This study introduces the concept of Critical Trajectory Points (CTPs) and establishes a CTP planning model to guide merging CAVs at a freeway on-ramp under the mixed traffic environment. A quintic polynomial trajectory model is applied to describe the lane-changing (LC) process of a merging CAV. The moving trajectories of potential collision points and the minimum safety spacing prediction model are derived. Considering feasibility and safety, the CTP planning model is formulated as a nonlinear optimization model to optimize the CTPs for CAVs within the control zone. The objective of the model includes maximizing the average speed and minimizing the LC time of merging vehicles within the control zone. Genetic Algorithm is designed for solutions. CTP strategy is more applicable to CAV merging guidance than complete trajectory strategy, especially with the accessible technologies in the near future. Numerical studies validate the advantages of the proposed model from the perspective of system optimum, even at low penetration rates of CAVs.