A warehouse where 600 cameras monitor every movement of goods, a "brain interface" where the employee only needs to think about the forklift for it to roll into position: Is it all just science fiction? Apparently not – but rather a reality, as the Logistics Future Congress in Dortmund demonstrates.

The human-machine interface, in the form of a headband resembling headphones, needs a moment to adjust to the employee's brainwaves – and the employee then has to concentrate intensely on the desired interaction so that the sensors register it. But after a short learning phase, it works, explained Michael ten Hompel at the opening of the future congress organized by the Fraunhofer Institute for Material Flow and Logistics (IML).

Ten Hompel, who has significantly shaped developments over the past 20 years, has no doubt that warehouse technology – like many other areas of society – is moving towards a "digital continuum" in which physical presence and algorithms operating in the background will increasingly merge. The speed at which this – as well as Chat GPT – is taking hold surprised many of the approximately 480 participants who attended the congress.

In fact, the development of AI-driven systems is progressing at a pace that no one had anticipated until recently. "Even we were surprised by the sudden speed," says Markus Hacker, European representative of the Californian graphics processor and game console manufacturer NVIDIA.

Graphic: T-Systems

The technology that allows neural networks to interact with each other actually dates back to the 1960s. "But back then, we didn't have the computing power to implement all of that," says Hacker. So, relatively suddenly, the applications are now advancing rapidly. At least half of the conference participants, by their own admission, are already using Chat GPT. This, in turn, trains the user, who then has to ensure that this "coaching" has a solid foundation and doesn't unexpectedly turn against its creator.

There are now almost no automated warehouse projects that don't rely on a digital twin—a digitized representation within the warehouse management system—to track and monitor their hundreds, sometimes thousands, of mobile components and robots. Simulation models train the machines. And action is needed: In ten to fifteen years, says Hacker, we won't have any truck drivers left and will be entirely dependent on automated and autonomous solutions. Train drivers are also becoming scarce and—like airplane pilots once were—are already largely trained in simulators.

It's an ideal field of activity for the Open Logistics Foundation (OLF) to design fundamental models and basic patterns that can then be specified to meet the needs of individual users. Stefan Hohm, chief developer at OLF member and logistics service provider Dachser, explains that the 400 to 600 high-resolution cameras installed under the warehouse ceilings at locations in Unterschleissheim near Munich and near Heilbronn track the respective goods movements – and do not use image recognition to identify individual employees (whether they're making a face or chatting with a colleague). "The identity of each employee is pixelated," says Hohm.

According to ten Hompel, ensuring that this actually happens and doesn't lead to total social surveillance like in China or a work-goal-oriented points system à la George Orwell should be guaranteed by a moral code, which, in the form of a Kantian imperative, should underlie every machine-controlled complex. Hence the idea of ​​a basic model that can be freely used for individual development by interested "users," but which would have to take fundamental ethical considerations into account. 

 In the Enterprise Lab, which Dachser operates jointly with the IML (Institute for Material Flow and Logistics), around 30 employees are currently working on a "Future Terminal" to achieve greater efficiency in goods handling. A project is underway with the AGV (Automated Guided Vehicle) provider Agilox to map recurring transport processes as efficiently as possible in a digital twin. Special AI-based algorithms interpret the data captured every second by optical scanning units to automatically and immediately identify, locate, and, in the future, measure all packages. The model pallet warehouse "@ILO" (Advanced Indoor Localization and Operations) also uses a commercial security system called "Splunk," which prevents, for example, an unauthorized person from logging in "simultaneously" on two different continents and committing misuse. A predictive model called PAnDA One indicates the anticipated storage and handling capacities required. Special AI-based algorithms interpret all data captured by the optical scanning units every second in order to automatically and immediately identify, locate and, in the future, measure all packages.

The employees have now come to view the AGVs not as "employees" but as a cooperation opportunity, allowing them to concentrate on other, more advanced activities.

Christa Koenen, a board member at DB Schenker (and also a member of the Open Logistics Foundation), sees AI in global supply chain management as a catalyst for agile supply chains. However, she believes that "intelligence" won't always be required to solve simple problems. "We shouldn't use a sledgehammer to crack a nut," says Schenker's head of IT and data. 

The Open Logistics Foundation is clear that hundreds of engineers shouldn't be developing new interfaces every day. "We need to tackle the fundamentals together"—so that everyone can focus their efforts on meaningful developments. "We're no longer just spectators." In Dortmund—unavoidable, given that the Germany-France match is currently taking place in the stadium next door—he sees the analogy to football: "We have to get on the pitch!"

According to estimates, in Germany, where, according to Uwe Clausen, head of the Institute for Transport Logistics at the University of Dortmund and board member of the Fraunhofer Institute for Material Flow and Logistics, around €311 billion was generated in logistics in 2022, approximately €10.5 billion is spent on logistics research. However, the US online delivery giant Amazon is reportedly spending significantly more than this double-digit billion-euro figure.

Photos: klk.

A magnificent evening event at the Phoenix des Lumières, featuring an immersive digital exhibition and monumental projections of works by Gustav Klimt and Hundertwasser, during which a Digital Logistics Innovation Prize was awarded, crowned the logistics future congress. 

Klaus Koch

www.iml.fraunhofer.de