Multi-layer mapping-based autonomous forklift localization in an industrial environment

Abstract

In this paper a forklift (mobile robot) localization method in an industrial environment is presented. The presented method is based on an adaptive Monte Carlo localization using three laser range scanners, where two safety lasers are located at the ground level and one laser is located at the top level of the vehicle. Each level(layer) has its own independent map acquired by mapping algorithms prior to the localization. The proposed localization method is based on two independent localizations at each level (layer) and their fusion. Presented method is tested in the realistic industrial environment and compared to the marker-based localization that is accurate enough to be considered a ground truth.