Integrating Global Navigation Satellite System and Road-Marking Detection for Vehicle Localization in Urban Traffic

Abstract

This paper presents an accurate vehicle self-localization system for autonomous driving. The developed system integrates multiple onboard sensors, a Global Navigation Satellite System (GNSS) receiver, an inertial sensor, a speedometer, and an onboard monocular camera to achieve lane-level performance in an urban environment. GNSS positioning suffers from the effects of multipath and non-line-of-sight (NLOS) propagation in urban canyons. To reduce the effects of multipath and NLOS propagation, this paper proposes using an innovative differential GNSS positioning technique with the aid of a three-dimensional building map. The road marking on the road surface provides visual information for driving. Recognition of road markings could be a way for localization when the position of the road marking is available. This research used the recognition of lane markings and stop-line markings to reduce further the positioning error. The multiple-lane detection was developed for improving the positioning error along the lateral direction. Stop-line recognition was used for optimizing the longitudinal position around intersections. The developed system was tested in an urban environment. Results demonstrate that the proposed method can provide submeter accuracy with respect to the positioning error mean and a 95% correct lane rate.