Largely unnoticed by the public, industrial automation is currently undergoing a revolution that is about connecting the real and virtual worlds. With many of its sensor solutions, SICK is already leading the way. For the past two years, this has been exemplified by PALLOC:the AI-assisted robot guidance system currently used for depalletizing and, in the future, for palletizing too.
As the name suggests, PALLOC (PALlet content LOCalization) is the intelligent SICK solution for one of the greatest challenges in today’s circular economy: millions of fully loaded pallets that need to be unloaded quickly and flawlessly at their destination, ideally around the clock. Due to the skills shortage, suitably qualified personnel are becoming ever more scarce. On the robot side, there has also been a lack of AI integration. What is needed, therefore, is a pioneering automation solution with outstanding adaptability thanks to AI.
PALLOC closes this gap, as it can be incorporated into real factory floors using plug-and-play technology. Added to this is the current emergence of the industrial metaverse on the boundary between reality and virtuality, which is opening up completely new approaches to automation. One of the first areas set to benefit from this development is, without a doubt, robotics – with its extremely high demand for adaptable AI solutions.
Taking PALLOC as an example, SICK can therefore already indicate the possibilities opened up by the use of virtual worlds. Initially, it does so with a view to virtually representing and expanding this and other robotics solutions in customer applications; prospectively, however, the idea will also be to put corresponding applications into operation in advance virtually. Global Industry Manager Robotics Dominik Birkenmaier and Jan Jarvis, Head of Visualization, are doing pioneering work in both worlds with their respective teams and in close cooperation with the product developers involved.
In the real world, PALLOC already constitutes the ideal connection between the Visionary-S 3D snapshot camera from SICK and a preinstalled and pretrained neural network. The result is a uniquely compact robot guidance set that works in a matter of seconds as an end-of-arm solution. Either mounted on the robotic arm or in a stationary position above the pallet, the system recognizes a theoretically unlimited number of box versions on the pallet’s uppermost loading level at lightning speed.
It identifies the contours and transmits the positional coordinates, including height information, to the robot control system so that the boxes can be grabbed and relocated. We say “unlimited number,” because the neural network, which has already been trained on a broad spectrum of box types before leaving the factory, can be expanded to include any number of further box versions via a user-friendly AI tool suite. This probably makes PALLOC, which does not require a separate industrial PC thanks to its built-in ALU, one f the most future-proof localization and robot guidance systems of its kind.
If the technology is to maintain and extend its lead, the metaverse needs to be utilized. Its virtual environment is designed precisely for developing, testing, and training AI. Thanks to the simulations now possible in this virtual environment, specific applications can now already be validated before they are deployed in the real world. In order to continue offering its customers innovative solutions, SICK is now working with NVIDIA, one of the world’s leading AI computing providers, and using their simulation technologies, such as the Isaac Sim™ platform, which is based on NVIDIA Omniverse.
In the case of PALLOC, with its integrated deep learning, this means that the functionality of a solution can already be represented, and potentially also validated, as a simulation in environments similar to the actual application. Building on this, it will also be possible in the future to train images and examples, as well as develop assessment processes, prior to practical deployment. “For us, the metaverse is thefastest and most sustainable way to incorporate sensors such as PALLOC within new customer applications,” explains Birkenmaier.
However, the possibilities extend much further. SICK solutions can be presented to customers in a visually appealing way in the future factory environment, such as at trade fairs. Where there is specific interest, this technology could also be used to verify whether a specific robot model performs its tasks correctly with PALLOC in a simulated warehouse – before it is even deployed in the real world. Any parameter modifications required could be made quickly and directly, ensuring that subsequent depalletization runs smoothly as desired. For Jarvis, working with the metaverse is all about eliminating certain tasks and uncertainties in reality, as they can simply be tested virtually. “Who really wants to go the long way around in the real world when you can set the right course at the outset, especially when going the long way around costs considerable time and money?”
And the journey is still ongoing. If depalletizing is possible, so is palletizing. PALLOC is currently being prepared for this new task using laboratory virtualization. So-called “mixed-case palletizing” is a whole different ball game, with much higher precision and speed requirements. But Birkenmaier believes SICK is on the right track: “What looks like Tetris with real boxes on the screen offers huge added value for our customers, including thanks to PALLOC."
