Bio-inspired Robotic Handling of Heterogeneous Logistics Goods

Abstract

In this paper we present a project to develop a robotic dual manipulator for the handling of heterogeneous logistics goods. The uncertain nature of these goods prevents the use of standard portal robots employing typical gripping mechanisms. In the logistics domain, the process of unloading goods from a container presents challenges to an otherwise highly automated and efficiency-demanding area. The great diversity of the goods with respect to shape, size, and rigidness, from rigid boxes to deformable bags, makes it difficult to introduce robotic systems that assist in logistics tasks. The primary goal of this work is, therefore, to develop a system capable of flexibly handling of goods with different physical properties (primarily weight, shape, and rigidness). Our approach involves integrating into a single robotic system the work from several technological fields: the use of visual and tactile information, control of robot-environment contact forces, dexterous gripping, dual-arm manipulation, and a certain level of a-priori knowledge. We believe that the use of behaviour-based software architectures along with the use of biologically-inspired concepts for building a mechanical gripper and designing reaching and handling behaviours will result in a robust robotic manipulator system which can achieve high performance levels in manipulating a great diversity of objects.