Our researchers led by the FTE-FENE team published in Science
Our researchers led by the FTE-FENE team have been published in Science for their groundbreaking development of technology that gives robots and prosthetic hands a sense of touch.
Robotic technology is progressing fast, but there's still a big challenge: making robots and prosthetic hands feel things the way humans do. Imagine a robot that can hold something fragile without breaking it, just like a person can. Right now, robots and prosthetic hands lack adequate capability of tactile perception.
Now, researchers at Uppsala University and Karolinska Institutet, coordinated by Docent Zhibin Zhang, have found a way to give robots and prosthetic hands a sense of touch, as reported in a paper published in Science (DOI: 10.1126/science.adf3708) on 10th May 2024. They took inspiration from how our own tactile nervous system works and created an artificial tactile system that mimics them. This system uses electric pulses to carry tactile information and simulate the feeling of touch.
With innovative design by assistant professor Libo Chen, the system consists of an electronic skin with tactile sensors that can detect touch, a new electronic circuit with a set of artificial neurons that turn those signals into electric pulses. The electric pulses were then processed in a neural network to figure out what's being touched and grasped. This system can quickly figure out what objects are just by touching and grasping them. And even if some sensors get damaged, it can still keep working pretty well.
This breakthrough is huge. It means robots and prosthetic hands could become safer and more useful, because they'll be able to feel things just like we do. They could also become better at handling objects delicately, like humans can. And there are even possibilities for creating sensory devices to help with medical stuff, like tracking diseases or help people who've had strokes.
This achievement was possible thanks to the close collaboration between researchers at Division of Solid-State Electronics and A. Özçelikkale group at Division of Signals & Systems on data processing and machine learning based on back-propagation for classification. Contribution by B. Winblad’s group from the Department of Neurobiology, Care Sciences and Society, KI, on testing the motor nerve of mice shows the potential of the new technology in medical applications.
Read more
Link to article: https://www.science.org/doi/10.1126/science.adf3708
Title: Spike timing–based coding in neuromimetic tactile system enables dynamic object classification
DOI: 10.1126/science.adf3708
Authors: Libo Chen, Sanja Karilanova, Soumi Chaki, Chenyu Wen, Lisha Wang, Bengt Winblad, Shi-Li Zhang, Ayça Özçelikkale*, and Zhi-Bin Zhang*