Variable crate picking by robot and machine learning

The turnkey systems can be integrated to achieve maximum availability and productivity, including machine learning-supported differentiation between different types of small load carriers - including those recognized for the first time - reliable gripping even with position tolerances and position errors, detection of damaged crates, position recognition and barcode reading or sorting of mixed small load carriers.

The depalletizing systems can currently be equipped with robots from the manufacturers KUKA, ABB or FANUC. Polyplan provides robots with a load capacity of 150 to 210 kilograms including a "safety reserve". In order to be able to pick up the various types of small load carriers flexibly and safely, Polyplan equips each depalletizing robot with a special multifunctional gripper with servomotor-driven pick-up mechanism, which moves automatically depending on the crate type and orientation and adapts to the position of the gripping holes of the respective small load carrier. An integrated, distance-measuring 3D sensor system - supported by machine learning - is used to differentiate between different types of KLTs. This is installed in a static position above the depalletizing station, creates an updated 3D image of the stacking scheme of the top pallet layer after each gripping and determines the contour, height and position values as well as the orientation of each small load carrier.

In addition to KLT, the depalletizing systems from Polyplan can also handle cartons and packaging containers made of expanded polypropylene (EPP) or - on request - other materials. The system technology described can also be used for the automated loading of pallets.