When the factory itself leads the way

The researchers at BTU developed the system in collaboration with OTH Regensburg. Their work has now been published in the journal “Sensors & Transducers”.

A sat-nav recalculates the route as soon as a road is closed – entirely automatically, without any manual intervention. It is precisely this principle that researchers at the Brandenburg University of Technology Cottbus-Senftenberg (BTU) now want to test in a factory hall.

The idea was developed by BTU researchers Annamria Arnouk and Martin Behm (Chair of Communication Technology) and Christian Borck (Chair of AI-Supported Automation Technology), together with a team from the East Bavarian University of Applied Sciences Regensburg (OTH Regensburg). 

If a workbench breaks down, the factory reorganises itself

In many factories, small, driverless transport trolleys move autonomously through the hall, transporting components from one station to the next – much like a robot moving parcels through a warehouse. If something breaks down at a station, the trolley must be rerouted. Until now, this has usually meant that a technician has had to reprogramme the robot manually. This takes time and is cumbersome.

The research team wants to change that: in future, the robots are to find the new route themselves – just like a sat-nav that automatically recalculates its route when a diversion is encountered.

The factory’s blueprint becomes a map for the robots

Almost every building has a digital floor plan. The researchers’ programme reads this plan like a map and automatically generates a complete route for the robots. What used to have to be laboriously set up by hand is now done by the computer in a single step.

At the same time, a second system uses a camera to constantly monitor whether everything is in order at a workstation. It also supports on-site staff in their work via smart glasses or a tablet. If it detects a problem, it alerts those involved.

What’s new is that both components are now directly linked. If the monitoring system reports a problem, it immediately requests a new route from the mapping service. The affected robots then automatically take a different route – without any human intervention whatsoever. The researchers tested this on a production line where electric drives for aeroplanes are assembled.

A digital twin for assembly

This second system was developed at the BTU Cottbus-Senftenberg. It comes into play precisely where the robots deliver their components: at the workbench, where employees assemble the parts by hand – in other words, the point at which machines and humans finally swap roles.

The workbench has a digital replica

At the heart of the system is a digital twin of the workstation: every component and every assembly step exists once again in 3D on the computer – like a virtual replica of the workbench that moves in real time alongside the actual workstation. When a Staff Member reaches for a screw, her digital counterpart reaches for it at the same time on the computer. If anything changes at the real workstation, the digital replica automatically updates – the two worlds run in sync.

AI provides support just like a good colleague 

This is made possible by an artificial intelligence system that analyses the camera images from the workbench: it recognises the employees’ hand movements and detects which component they are currently picking up – much like an attentive colleague who closely follows what is happening at the workstation next door without intervening themselves. As a result, the system knows at all times which step is currently being carried out and alerts the operator as soon as something is assembled in the wrong order or in the wrong position. A silent observer thus becomes a sort of early-warning system that intercepts errors before they become permanent in the finished assembly.

A risk-free training environment

The digital twin serves not only for monitoring but also for training: new employees can first run through complicated assembly steps virtually before working on the actual component – much like a flight simulator, in which trainee pilots practise their first manoeuvres safely. A wrong move in the simulation environment costs nothing more than another attempt; on the actual component, the same mistake could prove costly or delay production. Errors therefore have no consequences there, and the training process is noticeably faster – an advantage that pays particular dividends in the case of complex, infrequent assembly steps, for which a lengthy training phase on the actual workpiece would otherwise be almost unavoidable.

It is only when both systems work together that the true value of the research becomes apparent: Whilst the navigation system ensures that components arrive reliably at the correct station despite disruptions, the assembly assistance system ensures that they are also fitted correctly there – and reports findings back to the factory as soon as something is amiss.

A system that grows with every factory

To ensure the solution works not just in a single factory, the researchers are relying on technology that is universally understood – much like a map format that any sat-nav can read, rather than just one from a specific brand. If a production hall is refurbished, it will in future suffice to update the digital blueprint. The new route instructions for the robots are then generated automatically – without anyone having to manually readjust the system.

What’s next

So far, the idea is a well-thought-out concept. Next, the researchers want to test it in practice – in the BTU production hall in Cottbus. The aim: ultimately, nobody should have to initiate any actions manually. If a station fails, the factory – just like a sat-nav – should reorient itself entirely on its own.

Publication

A. Arnouk, M. Behm, C. Borck, T. Hoffmann, F. Dietrich, M. Melzer, J. Dünnweber: “Automated Mission Planning for Mobile Robots Using Building Information Modelling (BIM) and Digital Twins”. Published in: *Sensors & Transducers*, Vol. 273, Issue 2, June 2026, pp. 39–49.