单车道环形交叉口联网自动驾驶汽车的轨迹优化模型

The Trajectory Optimization Model for Connected Automated Vehicles in Single-Lane Roundabout Intersections

摘要: 在传统的车辆行驶环境中，环形交叉口相比常规交叉口具有较少的冲突点，然而，由于车辆在入口处需要减速避让，通常会导致车辆延误，从而影响通行效率。网联自动驾驶汽车(CAV)的出现，为提高交通基础设施的效率提供了巨大的机遇。本文提出了一种单车道环形车道的两阶段轨迹规划模型，旨在通过网联自动化技术，提升现有环形交叉路口的通行能力。该两阶段模型通过双层优化进行公式化处理：上层使用混合整数线性规划模型，优化车辆在环形交叉口的到达时间；下层则根据最佳到达时间，优化车辆的行驶轨迹，并以减少油耗为目标进行优化。最后，本文在一个典型的四臂环形交叉口上进行了仿真评估，验证了所提出模型的可行性，并探讨了不同控制区长度对模型性能的影响。

Abstract: In traditional driving environments, roundabouts have fewer conflict points compared to conventional intersections. However, due to the need for vehicles to slow down and yield at the entrance, delays occur, leading to lower traffic efficiency. The emergence of connected and automated vehicles (CAVs) offers significant opportunities to improve the efficiency of transportation infrastructure. This paper proposes a two-stage trajectory planning model for a single-lane roundabout, aiming to enhance the traffic capacity of existing roundabout intersections with the help of connected automation. The two-stage model is formulated as a bi-level optimization problem: the upper level uses a mixed-integer linear programming model to optimize the arrival times of vehicles at the roundabout, while the lower level optimizes vehicle trajectories based on the optimal arrival times, with fuel consumption reduction as the objective. Finally, the proposed model is validated through simulation on a typical four-arm roundabout, and the impact of different control zone lengths on the model’s performance is discussed.