Adaptive robotics allows robotic systems to respond to changes in their environment and production process based on real-time data. This involves a convergence of factors including cameras, force/torque sensors, and AI-based methods.

Increasingly, this technology is about more than just identifying objects. Cameras and computer vision have now become an established part of many applications, while force/torque sensors are becoming ever more important. Sensors of this kind allow robots to detect forces during a process and therefore to handle tasks in which parts aren’t always at exactly the same position. In conjunction with AI, new process steps can also be learned faster – and robots can be configured for different situations more flexibly.

Where classical robotics reaches its limits

This flexibility is of particular interest when it comes to complex assembly tasks. Fraunhofer IIS in Dresden is carrying out research into intelligent sensors, robust electronics, and AI methods for adaptive robotic systems.

One example is »Fiddle«, a technology that uses AI for motion planning in contact-rich assembly tasks, allowing the automation of processes that are difficult to implement with classical robotics.

The search for suitable robotic systems often starts well before automation with a consideration of the production process in question. Particularly for companies that – so far – have little to no experience with robotics, a neutral process analysis can highlight where automation is sensible and practical from a technical perspective. With more complex tasks, potential solutions can then be explored under realistic conditions as part of feasibility studies and demonstrators. Practical training sessions also help to transfer knowledge relating to adaptive robotics, sensor integration, and AI into industrial applications.

What challenges and opportunities does adaptive robotics present?

At the trade fair, it also became clear that the path to more flexible robotics is not without its issues. Cheaper robotic hardware – particularly from China – can reduce obstacles to investment. Some companies, however, express reservations and put greater trust in European manufacturers. Humanoid robots were another key topic of discussion – albeit with a nuanced approach to their technical, economic, and legal limitations.

Development is therefore moving toward robotic systems that can adapt more quickly to new tasks and respond to changing production processes. Expertise in adaptive robotics could potentially also be relevant to new robotics platforms, such as humanoid systems.

The key question is therefore not only »How much robotics is possible?« but rather »For which tasks does flexible and adaptive robotics actually create added value?« The latter question is vital in terms of how AI and robotics can move from research into industrial practice.