Multi-agent High-speed Lane Detection Based on Deep Reinforcement Learning
The detection of relative driving areas,lane line detection,and speed estimation have become crucial compo-nents.Some detection methods only detect lane lines with rough bounding boxes,ignoring the shape of specific curved lanes.This paper introduces deep reinforcement learning into the lane detection model,using intelligent agent positioning point movement to achieve accurate lane detection and positioning.To improve the auxiliary driving system and improve the recognition of lane markings.Both the driver and the auto drive system can improve the driving experience.
deep learningreinforcement learninglane line data detectionintelligent agentsreward mechanisms
万方数据
维普
