The human brain, a marvel of nature, is the cornerstone of our central nervous system. With its intricate network of around 85 billion neurons, it orchestrates our cognitive functions, sending and receiving signals that govern our bodily processes. These neurons, the building blocks of our nervous system, transmit information, allowing our bodies to coordinate and function harmoniously. The precision of these interactions has inspired an innovative robotics experiment, known as "neurocellular automata."
"In simple terms, it's a computing model inspired by the brain's neuronal structures. It's like a cluster of neurons communicating and informing each other on how to act, which in turn determines the robot's local operations," explains Kasper Støy, a robotics professor and member of the REAL research group at the IT University of Copenhagen.
Støy is the scientific coordinator of MOZART, an EU initiative focused on developing reconfigurable surfaces with soft sensors and AI-powered learning tools. One of their key partners is the Italian Institute of Technology (IIT) in Genoa, led by Lucia Beccai, who specializes in soft biorobotics perception. They are working on integrating these technologies into a movable platform, with a focus on handling delicate objects.
"Imagine a next-generation conveyor belt, but in 3D. We're adding a third dimension to the motion, allowing objects to be flipped and rotated, not just moved from point A to B. The surface adapts to the object's shape, manipulating it gently," Støy describes. This sophisticated solution consists of around ten modules, each functioning as an individual robot, creating a soft membrane that can lift and change shape.
"It's like a checkerboard of small robots working together. Each element has its own neural structure, communicating with the surrounding elements, just like the human brain. This coordination is key to the success of the system," Støy emphasizes.
But here's where it gets interesting: the system can learn and adapt its behavior through machine learning. It can understand the weight, softness, and fragility of objects, making it extremely versatile. "Instead of hard-coding, we let the robot learn through exposure to various variables. This knowledge is then encoded in the neural network," Støy explains.
The potential impact of this technology is vast. It could automate industrial sectors, reducing costs and making products more affordable. "Many products today are affordable due to automated production chains. By introducing this technology, we could reduce the price of food, making it accessible to more people," Støy suggests.
Current applications are focused on the food-packaging sector, particularly for delicate items like tuna and poultry. "Soft robotics can ensure gentle, safe interactions, unlike rigid manipulators that might damage the food. It's a game-changer for demanding packaging processes," Beccai adds.
MOZART, now in its final year, is moving into the practical phase, with the first demonstration expected soon. "The AUTOMAT manipulation surface is designed for 'chicken-sized' objects, but the concept can be scaled up or down for various applications," Støy notes.
The possibilities are endless. From assisting the elderly and disabled to handling dangerous tissues and cells, gentle soft robotics has the potential to revolutionize many fields. "While we need to improve biocompatibility, the applications are vast and exciting," Beccai concludes.
