Research On The Method Of Autonomous Vehicles Passing Through Unsignalized Intersections

Autonomous vehicles（AVs）have the functions of environmental perception,precise positioning,path planning,and intelligent control.With the assistance of V2X technology,autonomous vehicles can communicate and cooperate with each other,and can pass safely and quickly at intersections without the assistance of signal lights,greatly improving the traffic efficiency of intersections.In order to ensure the safety of vehicles at unsignalized intersections and reduce vehicle delays,this paper proposes a spatial-temporal trajectory distributed planning method for multiple autonomous vehicles with the goal of optimal single vehicle speed.The spatial-temporal trajectory information of the planned vehicle is used as the planning condition of the vehicle to be planned,and the optimal safety speed is solved for the vehicle to be planned.Firstly,a buffer speed control model based on the car-following model is established to determine the feasible speed range of the planned vehicle in the buffer zone.Secondly,the physical space inside the intersection is gridded,and the grid through which the vehicle to be planned passes is determined by the driving path of the vehicle to be planned.According to the time reservation information of the completed planned vehicle on the grid,the time and space resource constraints on the optimal safety speed of the vehicle are carried out,that is,when the planned vehicle passes the intersection at the optimal safety speed,the time interval of entering and leaving any grid on the path cannot intersect with the occupied time resources on the grid.Finally,the optimal safe speed of the vehicle to be planned is solved by buffer speed constraint,conflict zone speed constraint and road speed limit constraint.In the intersection,four different sizes of test vehicle groups are simulated in buffers with different lengths.The experimental results show that when the buffer length cannot meet the vehicle speed regulation requirements,the average vehicle speed decreases and the average delay increases significantly.At this time,increasing the buffer length can quickly reduce the average vehicle delay.The buffer length has limited ability to reduce the average delay of vehicles.After reaching the peak,the average delay of vehicles will have a significant positive correlation with the buffer length.Compared with the signalized intersection,the spatio-temporal trajectory model proposed in this paper improves the throughput of the intersection,and is less affected by the traffic density of the intersection,and has higher stability.At the intersections with traffic flow of 600pcu·h^-1,1200 pcu·h^-1and 1800pcu·h^-1,the average driving time of vehicles is reduced by 67.41%,70.94%and74.70%respectively.In order to realize the free steering of the lane,the space-time trajectory model of the unsignalized intersection is optimized,and the simulation experiment is carried out in the T-intersection.The simulation experiment verifies the feasibility of the model at the T-intersection.By comparing with the signalized intersection,it is confirmed that the model is superior to the signalized intersection model and effectively reduces the average driving time of the vehicle at the intersection.