Optimal smoothing based mapping process of road surface marking in urban canyon environment

This paper proposes a road surface marking (RSM) mapping process based on an optimal smoothing method using a probing vehicle equipped with high accuracy sensors for the localization of an autonomous vehicle in regions experiencing GPS outages. Since the RSMs in the map can be inferred by the trajectory of the probing vehicle, it is important to estimate the precise trajectory for precise RSM mapping. For the trajectory estimation in GPS outage regions, an optimal smoothing algorithm is applied to the mapping process. The algorithm can estimate trajectories more precisely by integrating future measurements as well as past and present measurements. The RSMs can be estimated by point clouds measured by light detection and ranging (LIDAR) through deskewing, ground extraction, intensity calibration, and data mapping along the trajectory. Finally, the RSM mapping process was evaluated in an experiment on Samsung Street, Seoul, South Korea, which is a high traffic-area with many skyscrapers.