Expert System For Flexible Palletizing Of Mixed Size And Weight Parcels

Abstract

Stacking boxes of mixed size and weight is a tedious task requiring intelligence. It requires the ability to recognize that boxes are available to be stacked, the recognition of different box types, the selection of a grasping point for picking up the box and most importantly, a determination of where to stack the box on a partially loaded pallet. The purpose of this paper is to describe an expert system for determining how to stack a set of mixed size and weight boxes and provide control information to a robot to perform actual palletizing. A prototype system has been developed using an expert programming language and an industrial robot work cell. The system has been tested with actual food parcels weighing up to 50 pounds and performs very well. The formation of flat regions is one of the many rules implemented in the expert system. The system also works at speeds comparable to human speeds. The study demonstrates, in large measure, the feasibility of the use of expert system AI techniques and industrial robots for palletizing mixed size weight parcels in a general workhousing application.