Research On Visual Positioning And Speed Guidance Of Autonomous Vehicle At Intersection

With the rapid development of automatic driving technology,Intelligent and Connected Vehicles have also been continuously improved with the development of communication technology.At present,there are more and more vehicles with autonomous driving functions,and there are more mixed vehicles in urban traffic,so the traffic operation is more complex than the traditional traffic flow.In addition,the autonomous vehicle perception system is equipped with various sensors,radar,onboard cameras,etc.They can sense a variety of road conditions.The image information acquired by the vehicle camera is more abundant so that it can be explored for more applications.At present,intelligent and connected vehicles are gradually increasing,and urban transportation will face a more complex traffic proposition of mixing autonomous vehicles and traditional vehicles.In this case,the autonomous vehicle needs to tap its own sensing technology,using the on-board video to detect the stop line,so as to guide the speed of the traffic flow at the intersection,making the traffic flow at the intersection more efficient.This study focuses on the processing of on-board video,detecting the stop line on the basis of the current image processing technology,focusing on the combination of image processing and the scene of the on-board video,and using traditional image features to identify the relative position of the stop line in the case of missing scene image sets.Based on the theory of vehicle safety distance and the idea of hybrid modeling,this method proposes a speed guidance method for mixed traffic flow at intersections.This method not only considers the different tracking states of mixed traffic flow,but also considers the following characteristics of two vehicles,distinguishes different driving states for different types of vehicles,and finally combines the relative position information obtained by image processing to guide the speed of intersection vehicles to achieve the effect of the safe and fast passage of vehicles.After verification,the current detection of the stop line can obtain about 66%recognition accuracy,the speed guidance strategy has a significant effect at the intersection,and the efficiency of intersection traffic is significantly improved.The consequence shows that the speed guidance method considering the real-time detection of the intersection stop line in this study has an effect on improving the traffic capacity at the intersection.The state of the intelligent network connection will not be fully realized in the short term,and there is a lack of speed guidance methods for the mixed traffic flow of the circulation from traditional traffic to an autonomous vehicle.This study proposes solutions to intersection scenarios in this context,so that traffic control and guidance strategies are more in line with the development of traffic and the actual needs.