Forum on the integration of AI and cobots: Experts discuss future developments at the Second Swiss Robotics Forum in Biel/Bienne

Artificial intelligence (AI) currently has a dominant presence in media headlines – albeit mostly in connection with ChatGPT and other chatbots. What is not attracting anywhere near as much attention is the combining of AI with other applications. Around 35 experts from universities and industry met on the 23 May 2024 at Switzerland Innovation Park Biel/Bienne for an innovative forum organised jointly by SUPSI, the Swiss Cobotics Competence Center (S3C) and SATW. Led by Anna Valente, who is a professor at SUPSI and head of the SATW Intelligent Manufacturing Topical Platform, the event thematised the integration of AI and collaborative robotics (cobotics) in various sectors.

Anna Valente emphasised the importance and potential of such integration, but also pointed out that its acceptance in industry is still rather low. The aim of the event was to shed light on short-, medium- and long-term developments with benefits for industry.

Short-term developments

Ambra Vandone from SUPSI presented a project called ‘The shining cobotic cell’. This focusses on robots’ handling and inspection of geometrically complex shiny objects, such as watch cases. AI-based image analysis with advanced vision systems and specialised robotic arms are essential for this, even though human operators will continue to play a central role in the evaluation of detected defects in the future.

In addition, Mike Hausdorf from Siemens presented advances in so-called ‘piece picking’, demonstrating how, with sophisticated software and AI, robotic arms can handle objects of very different shapes and sizes. This technology, which simplifies tasks ranging from conveyor-belt work to logistics, offers promising opportunities for collaboration between cobots and human workers.

Medium-term developments

Mohamed Bouri from EPFL presented innovations in haptic interfaces that go beyond the hands to also include the feet. This enhancement enables scenarios similar to playing an organ and allows four cobots to be operated at the same time, which is particularly advantageous for medical applications. However, putting these advances into practice requires extensive machine learning to translate human movements into precise cobot actions. The human capacity for manipulating three or four things at once is currently being researched.

Long-term developments

Looking further into the future, Stelian Coros from ETH Zurich introduced the concept of ‘physical AI’: The potential of AI will be tapped to create intelligent machines that can perceive the physical world, make decisions and act accordingly. For such development, there are challenges to overcome, such as perceiving the environment in an abstracted yet precise manner and describing the objects to be handled. Digital twins could prove helpful here.

The forum served as a platform for inspiration, innovation and collaboration, while highlighting current efforts to integrate robotics and AI in various sectors. These developing technologies have enormous potential – and promise new opportunities, not only in industry, but also in medicine.

Impressions