Strategies for 3D data acquisition and mapping in large-scale modern warehouses

Abstract

Today modern manufacturing and logistic processes rely in most instances on manual forklifts. Being known for low efficiency, high energy consumption and listed among the most frequent causes of severe accidents in factories, the SICK AG coordinated EU funded research project PAN Robots aims to replace them in large scale with Automated Guided Vehicles (AGV). Major obstacles for operators with current AGVs are high initial costs, e.g. registration of all pallet positions and reflectors which needs to be performed by highly trained professionals in order to create an accurate navigation map for the AGVs. Simultaneous Localization And Mapping (SLAM) is the scientific approach to initial mapping. In this paper, we present 3D mapping strategies dedicated for the warehouse environment with multiple Laserscanners. Preliminary results are shown based on measurement data captured at the site of our project partner Casbega, Madrid, Spain.