A Research Of Traffic Cone Lane Recognition And Path Planning In Autopilot

With the development of the automobile industry,the number of cars at home and abroad sustain growth,which is convenient for people to travel also cause some negative effects.The automatic driving technology helps to reduce the rate of traffic accident result from some operational errors such as fatigue driving and inattention.it will greatly improve the road transportation efficiency and save travel time conbine with the V2 X technology.Autopilot in different scenes have different requirement,in oder to adjust the specific scene,autopilot needs a great improvement according to the actual environment.Expressways and urban roads are often maintained and repaired.Traffic cones are generally used to isolate working areas and build temporary driveways,there is currently no good way to deal with such road conditions.Therefore,this paper focuses on the special scene of temporary roads which built by traffic cones,researching the target recognition methods and path planning and tracking ways of cone lane.The work of this paper is summarized as follows:(1)Road cone and lane line recognition based on vision.Firstly,put forward a method of traffic cone recognition based on color feature segmentation after the image processing.Secondly,calculate the location of traffic cone.and then recognize the lane line based on Canny edge detection and Hough transforming.Finally,fit the lane line equation by the least square method to complete the task of target recognition.(2)Path planning and tracking.Firstly,fit the cone lane line on the basis of target recognition,then judge the feasible driving area of the vehicle combine the lane line and calculate a target path.Secondly,establish the coordinate system and the vehicle kinematics model,then derive their transformation relationship.Finally,track the path by using the control method based on visual heading deviation feedback(3)Establishment of experimental platform.Based on the in-depth investigation of the existing automatic driving technology,this paper analyzes and puts forward the methods in cone Lane scene.Design the hardware scheme of the experimental platform,build and test the complete hardware platform,formulate the corresponding underlying control and execution software scheme,and lay the foundation for subsequent algorithm verification.Finally,the vision-based target recognition algorithm and path planning algorithm are verified,and images of various working conditions that may appear in the target scene are taken.The test results show that the road cone recognition algorithm proposed in this paper can achieve good results under different lighting conditions.The recognition effect of the lane line can accurately detect the lane line in different road conditions.The actual vehicle test results prove that the path planning and path tracking algorithms in this paper are feasible.