Automatic unloading of heavy sacks from containers

Automated unloading of mass goods is of great importance for cost-effective logistic processes. Furthermore, companies are interested in promoting the automation of their logistic processes. In this paper the unloading of heavy, inflexible sacks from containers is addressed. The sacks are stacked neatly in rows of piles. Our approach combines object recognition and object handling while taking their dependencies into account. For 3D scene data acquisition a 2D laser scanner, which is mounted on a pitch device, is employed. The object recognition segments the acquired data by taking advantage of the sack surface characteristics. For sack handling a gripper consisting of a system with extendable and fixed claws is presented. Our system is the result of a feasibility study showing that sacks which are heavy and resistant to bending sacks can be handled automatically. Therefore further research is necessary to build a prototype system.