Collision Avoidance Algorithm Of Continuous Flow Intersection Under Autopilot Background

In order to improve the capacity of traditional intersections,many scholars try to change the rules of time and space of vehicles running at intersections.Aiming at the conflict traffic flow of left-turn direct traffic,many traffic organization and signal schemes are put forward,such as far-cause turn,left-turn of exit lane,U-turn far-lead of left-turn,bow-tie intersection and so on.Among them,there are two main signal control schemes: Continuous Flow Intersection(CFI)and tandem intersection(TI).CFI is widely used all over the world.It has remarkable effects in improving the capacity of intersection,shortening signal cycle,improving the utilization of road resources and reducing delay.However,traffic conflicts in its operation mechanism lead to the design of at least one right-turn lane for right-turn vehicles.This enlarges the area of intersections and is not conducive to the secondary transformation of the original urban intersections,resulting in their inappropriateness to small intersections.The rapid development of automatic driving technology will,to a certain extent,peak the existing traffic control theory.V2 V,V2I and even V2 X technology can help vehicles to share running status information with each other in the course of driving.If vehicles pass through intersections according to certain traffic rules,the computer will be able to predict the possible dangers in the future,so as to control the vehicle in advance to avoid.At present,the existing car-following models take the current motion state of the front vehicle as the target of car-following,which has nothing to do with the running state of the rear vehicle,and consider that the vehicle is only affected by the current traffic state and has nothing to do with its future running state.The existing car-following model can guarantee the safe distance between front and rear vehicles,but under the conditions of diversion,confluence,interleaving and so on,the optimal running state can not be obtained.At present,there is still a lack of collision avoidance model for the future traffic state aiming at the perceptual characteristics of autonomous driving vehicles.In this paper,a CFI intersection without right-turn lane is designed,and a new right-turn vehicle trajectory is designed.Three potential collisions at CFI intersection are analyzed: direct-right turn shunt,direct right turn confluence and right-shift left turn conflict.A new control strategy is proposed to make the self-driving vehicle follow the future traffic state.The section where the avoidance behavior occurs is transferred from the curved section to the straight section,which improves the ride comfort of passengers.For different collision situations,FIFO is revised to arrange vehicle traffic sequence reasonably,which ensures vehicle safety.According to the control strategy,a collision avoidance model aiming at the future traffic state is established through kinematics.Select reasonable input parameters,calculate the control speed,acceleration,control time and other output parameters of the collision avoidance model,and compare with the traditional intersection to analyze the optimization effect.The results show that the collision avoidance model can improve the speed of vehicles leaving the intersection,reduce traffic conflicts to a certain extent,ensure vehicle safety and improve ride comfort.Combining with the characteristics of state perception and information sharing,the collision avoidance model proposed in this paper is an improvement of the traditional safe distance model.It can reduce the right-turn lane at the left-turn intersection,optimize the speed and acceleration of the vehicle at the intersection,and also provide a new idea for the traffic rules of the automatic driving vehicle at the intersection.