Analysis Of Driving Scene And Path Planning Of Intelligent Vehicle At Left-turn-waiting Intersections On Urban Roads

Intelligent vehicle are experiencing the development stage of coexistence with humandriven cars,and it is challenging to solve the dynamic path planning problem under the driving scenario of left-turn-waiting intersections on urban roads to ensure the safe and efficient driving of autonomous vehicle.Intelligent vehicle is suposed to pass through left-turn-waiting intersections safely.In order to achieve the goal,this paper analyzes the driving scenario of double left-turn-waiting intersections on urban roads,and studies the corresponding path planning algorithm,mainly includes: 1.Investigating and studying the road intersections containing left-turn-waiting lanes in the main traffic artery area of the case city(Chengdu),designing a data collection scheme based on video filming,and collected key data of the traffic of left-turn vehicles at different time pieces at the case intersections.2.Statistical analyzing the target lane selection characteristics of left-turn vehicles at double left-turn-waiting intersection,and analyzing its trajectory and speed characteristics by extracting the kinematic information of vehicle traffic through video image processing.3.Under the conditions of interfering vehicles based on the kinematic characteristics of left-turning vehicles,analyzing the target lane selection combinations of intelligent vehicles,constructing double left-turning-waiting intersection driving scenario,and studying the safe driving strategies of intelligent vehicles in this scenario.4.Establishing quintic polynomial local path planning algorithm model based on Frenet coordinate system,and using the algorithm to simulate and verify three types of typical scenarios.The innovation points of this paper: 1.The scenario at left-turn-waiting intersection of intelligent vehicles driving was found and studied,which enriched the automatic driving scenario library.2.A local path planning algorithm was presented,which is for the intelligent vehicles driving scenario at left-turn-waiting intersection.At double left-turn-waiting intersections,the research in this paper shows that driving scenarios can be three categories: conflict-free,merging conflict and cross-conflict.The simulation shows that quintic polynomial local path planning algorithm based on Frenet coordinate system can plan a safe and smooth driving path.So the autonomous vehicle can pass the double left-turn-waiting intersection with stable speed in the conflict-free scenario,and can avoid driving conflicts in the merging conflict and cross-conflict scenarios.