人机混驾交通流交织区换道模型切换控制策略

因交织区的强制换道存在紧迫性, 车辆换道行为在交织区后半段会出现因换道意愿强烈而产生的激进换道行为, 这种微观的换道行为将给交通流带来一定影响; 在人机混驾情形下, 不同类型换道切换控制模型同样可能影响交织区通行能力。在分析人机混驾交通流交织区换道行为特性的基础上, 将换道类型分为保守型换道和激进型换道; 在可接受安全间隙模型的基础上结合自动驾驶车辆间的协同行为, 构建自动驾驶车辆在保守状态下的协同换道模型; 以及在激进型状态下考虑目标车道后车类型影响下, 构建激进型换道模型。通过分析津保立交桥实地调研轨迹数据和NGSIM中US-101交织路段轨迹数据, 分别拟合了保守型、激进型换道模型切换点分布函数; 考虑不同车辆驾驶行为特性及其相互作用, 提出人机混驾条件下换道模型切换控制逻辑决策。以SUMO仿真软件搭建实验平台, 考虑人工驾驶车辆换道模型切换点分布特性, 以优化最大流率、交织区整体车辆运行速度、换道车辆速度等为目标, 确定不同自动驾驶车辆渗透率下自动驾驶车辆的最佳保守型-激进型换道模型切换点。仿真结果显示: 在交织区长度为250 m, 自动驾驶渗透率分别为0.2, 0.5, 0.8时, 自动驾驶换道模型切换点分别在180, 80, 50 m处达到最佳, 即随着自动驾驶渗透率的提高, 换道切换点最佳位置将向交织区入口处逐渐移动, 且在自动驾驶渗透率较低时这种换道切换点的变化较为明显; 在较高渗透率下, 由于协同换道出现频率增高, 自动驾驶强制性换道行为比例降低, 换道模型切换点对交织区通行能力的影响逐渐变小。本项研究对人机混驾条件下高速公路交织区自动驾驶车辆的换道控制提供决策依据 

Switching Control Decision of Lane-changing Model in Interweaving Areas of Mixed Traffic Flow with Human-driving and Autonomous Vehicles

Due to the urgency of forced lane change in weaving areas, lane-changing behaviors occur in the second half of a weaving segment due to a strong desire to change lanes, which will have a certain impact on traffic flow. In a situation of mixed traffic flow with human-driving and autonomous vehicles, different lane change control models can affect the capacity of weaving areas. Based on analyze the characteristics of lane-changing behaviors in the weaving areas with the mixed traffic flow, they are divided into two types: conservative lane-changing and radical lane-changing. Based on an acceptable safety gap model and the cooperative behavior among autonomous vehicles, a cooperative lane changing model for autonomous vehicles in a conservative state is constructed; and the radical lane change model under the influence of the vehicle type behind the target lane in the radical state. By analyzing the track data from the field survey of Jinbao Interchange and the track data of the US-101 weaving area in NGSIM, the distribution functions of switching points of conservative and radical lane changing models are fitted, respectively; Considering the characteristics of different vehicle driving behaviors and their interactions, the logic decision of lane change model switching control under the condition of the mixed traffic flow is proposed. The SUMO simulation software is used to develop an experimental platform. Considering the distribution characteristics of the switching points of the lane-changing model of the manual vehicles, and aiming at optimizing the maximum flow rate, the overall vehicle running speed in the weaving area, and the speed of the lane-changing vehicles, the optimal conservative-aggressive lane changing model switching points of the autonomous vehicles under different penetration rates of the autonomous vehicles are determined. The simulation results show that when the length of the weaving area is 250 m and the penetration rate of autonomous vehicles is 0.2, 0.5, 0.8, the switching point of automatic lane-changing model reach the best at 180, 80, and 50 m respectively, with the increase of the penetration rate of autonomous vehicles, the best position of the lane change switching point will gradually move towards the entrance of the weaving segment, and the change of this lane change switching point is more obvious when the penetration rate of autonomous vehicles is low; At higher permeability, due to the increased frequency of cooperative lane-changing, the proportion of autonomous vehicle forced lane changing behavior decreases, and the impact of lane-changing model switching points on the capacity of weaving area gradually decreases. This study provides a basis for lane change control decisions of autonomous vehicles in freeway weaving areas under the condition of mixed traffic flow.