基于冲突相位组的自动驾驶交叉口行人过街控制方法

Pedestrian crossing control method at autonomous intersection based on the conflict phase groups

【目的】在无信号控制的自动驾驶环境下,自动驾驶车辆的通行轨迹将与过街行人产生大量冲突,如何利用交通控制手段使行人安全通过交叉口,并避免对自动驾驶车辆的通行造成较大的干扰,是亟待解决的关键问题。【方法】本文提出一种基于冲突相位组的自动驾驶交叉口行人过街控制方法,将到达交叉口的车辆流向分为4个冲突相位组,在各相位组内单独分配通行时间,基于冲突相位组对自动驾驶车辆和行人过街的通行时间进行建模;在穿插式通行模式的基础上,使用行人信号灯保障行人过街需求,建立考虑行人二次过街的自动驾驶交叉口交通控制模型。模型以交叉口各流向需求量与实际交通量乘积之和最大为目标,以各流向允许车辆通行的时间比例和行人信号灯状态为决策变量,综合考虑交通流量、行人和车辆通行权等约束,建立混合整数线性规划模型(mixed-integer linear program,MILP),该控制模型可为各流向的车辆和行人分配通行权。【结果】本文模型的车均延误较定时控制方案的降低26.74%,较单次过街模型的降低11.53%,人均延误较定时控制方案的降低51.66%,较单次过街模型的降低36.20%。这表明本文模型能有效提升交叉口的通行效率。【结论】本文模型能根据自动驾驶车辆和行人的通行需求,对交叉口时空通行权进行分配,有效保障行人过街安全。

[Purposes]In the automatic driving environment without signal control,there will be a large number of conflicts between autonomous vehicles and pedestrians.It's a critical issue to be addressed in how to ensure the safety of pedestrians'crossing at autonomous intersections and avoid causing greater interference to autonomous vehicles.[Methods]This paper proposes an unique traffic control method for pedestrians'crossing based on the conflict phase groups.The vehicles approaching the intersection are divided into four conflict phase groups according to their movements,and the passage time is allocated separately within each phase group.The passage time of vehicles traversing the intersection are then modeled based on the phase groups.Based on interspersed traffic mode,traffic lights are utilized to ensure the safety of pedestrians'crossing,the two-stage crossing for pedestrians is also employed to improve the overall efficiency of the intersection.A mixed-integer linear programming(MILP)model is formulated,aiming to maximize the sum of the product of demand and actual traffic volume for each direction at intersections.The decision variables of the model include the time proportions for vehicle movement in each direction and the pedestrian signal states.Con-straints are incorporated to comprehensively consider factors such as traffic flow,pedestrian rights of way,and vehicle passage rights.The allocation of right-of-way for both vehicles and pedestrians in each direction can be determined by this model.[Findings]The average delay of vehicles in this model is reduced by 26.74%compared with the fixed plan scheme,11.53%compared with the one-stage crossing model,and the average delay of pedestrian is reduced by 51.66%compared with the fixed plan scheme and 36.20%compared with the one-stage crossing model.The result clearly shows that the proposed model can effectively improve the traffic efficiency of intersection.[Conclusions]The proposed model can allocate the space-time right of way of intersection according to the traffic demand of autonomous vehicles and pedestrians,and effectively guarantee the safety of pedestrians.