自动驾驶环境下面向交叉口自由转向车道的交通控制模型

Intersection Traffic Control Model for Free-turning Lane in Connected and Autonomous Vehicle Environment

在未来自动驾驶环境下,自动驾驶车辆之间能相互配合、相互穿插地通过交叉口,而无需信号灯控制。因此,有必要研究新一代的能保障自动驾驶车辆安全高效通行的交叉口控制模型。已有控制模型可分为基于交叉口空间离散的控制模型和基于交叉口冲突点分析的控制模型,目前主要存在控制方式和模型非线性等方面的不足。建立了基于混合整数线性规划(MILP)的自动驾驶交叉口控制(Autonomous Intersection Control,AIC)模型,设计交叉口自由转向车道,允许交叉口所有进口道都能"左直右"通行,将交叉口空间离散为等距网格并建立网格坐标方程,考虑车辆在交叉口内部的行驶轨迹,建立车辆轨迹的上边界和下边界方程,确定行驶轨迹压过的交叉口网格,并建立网格被车辆路径占用的时间方程,使用同一网格同一时间只能被一台车辆占用的冲突点约束保障交叉口安全通行。模型以所有车辆通过交叉口的总延误最低为目标函数,通过将约束条件线性化处理,使用AMPL(A Mathematical Programming Language)并调用Gurobi数学规划优化器对模型进行求解。最后对模型效益进行了案例分析。结果表明:所提模型能有效处理自由转向车道的交通流到达模式,对比已有模型经常采用的先到先服务控制策略,该模型能整体优化车辆通行方案,降低车均延误50.51%,降低最大车辆延误29.12%,同时交叉口空间利用率提高了66.17%。

In the future automatic driving environment,autonomous vehicles can cooperate with each other and cross intersections without signal control.Therefore,it is necessary to develop a new type of intersection control model which can ensure the safe and efficient movement for autonomous vehicles.Existing control models can be divided into two types.The first one is based on discretizing intersection space.The second is based on conflict point analysis.Current studies on these two models often adopt simplified control strategies and solving these models can be inefficient owing to the nonlinearity of these models.This paper established an autonomous intersection control model based on mixed-integer linear programming,in which free-turning lanes were designed at the intersection to allow traffic to turn left,go straight,and turn right.In particular,the grid coordinate equations were first established by discretizing the intersection space into equidistant grids.Secondly,by considering the vehicle trajectory inside the intersection,the upper and lower boundary equations of the vehicle trajectory were established to determine the intersection grids to be traversed by the trajectory,and then the duration that the grids are occupied by the vehicle trajectory can be calculated.The constraint that a grid can only be occupied by one vehicle at a time was used to ensure the safety of the intersection.The total delay of all vehicles travelling through the intersection was adopted as the objective to be minimized.By linearizing the constraints,A Mathematical Programming Language software and Gurobi mathematical programming optimizer were used to solve the model.Finally,the benefits of the model were analyzed.The results show that the model can effectively capture the traffic arrival in free-turning lanes.Compared with the first-come-first-serve control strategy which is commonly used in existing models,the proposed model can optimize the whole vehicle trajectory plans,effectively reduce the average vehicle delay by 50.51%,reduce the maximum vehicle delay by 29.12%,and increase the space utilization of the intersection by 66.17%.