Implementation of autonomous driving of a ground vehicle for narrow high-curvature roads using surround view images

Abstract

In this paper, a new approach for the implementation of autonomous driving of a ground vehicle on narrow and high-curvature roads using a surround view system is proposed. The approach called the independent frame consecutive searching technique is an algorithm which is developed to detect the left and right lane markers of a lane stably and to generate virtual lane marker lines especially when the roads are narrow and have sudden corners and windings with some faded lane markers. Then the steering angle of an autonomous vehicle is calculated so that the vehicle drives along the desired path after deriving the smoothed virtual centerline. In the real-time experiment, the suggested approach achieved robust detection of lane markers using surround view images and drove a test vehicle autonomously with the generated desired paths on such narrow and high-curvature roads with any lane departures. Therefore, this study shows that the method can be applied to the development of autonomous vehicles.