Experts developed an e-skin that can decode complex human motions using enhanced deep-learning technology. According to Tech Xplore's latest report, a sensor that can act as an electronic skin was recently developed by the researchers at Korea Advanced Institute of Science and Technology (KAIST) and Seoul National University, integrating it with a deep neural network.

Also Read: MIT Experts Develop Washable Coronavirus Sensors That Can Be Embedded In Clothing

The newly developed e-skin system can capture human dynamic motions, including rapid finger movements from a distance using enhanced deep-learning technology; the innovation was presented in a paper published in Nature Communications. Experts in the fields of computer science and mechanical engineering had an interdisciplinary collaboration that gave way for the development of the new system. Sungho Jo, a computing professor at KAIST, and Seung Hwan Ko, a professor of mechanical engineering at Soul National University, lead the new study.

ALSO READ: NHS to Conduct a Virtual Reality Trial with Hi-Tech Headset to Tackle Patients' Worst Phobias

Professor Seung Hwan Ko has been generating cracks in metal nanoparticle films using laser technology over the past years in the hope of developing highly sensitive strain sensors. The researchers integrated the resulting sensor arrays to a virtual reality (VR) glove specifically designed to identify human movements.

How can skin sensors capture complex human motions using deep-learning technology?

According to Tech Xplore, researchers developed an electronic skin sensor that decoded complex human motions, such as rapid finger movements, using enhanced deep-learning technology. Professor Seung Hwan Ko said the complexity of the target system increases. The required number of strain sensors also increases, which is why his lab typically uses five to ten strain sensors to predict complex hand motion; each finger must be integrated with at least one or two sensors.

"A few years ago, I started asking myself the following question: Can we accurately predict hand motion with only one single strain sensor instead of using many sensors? Initially, this appeared to be a dumb question, because it was almost impossible to tell what finger the signal from a strain sensor came from," he further explained.

On the other hand, Professor Sungho Jo was creating different strategies to integrate machine learning techniques with state-of-the-art sensors. Professor Ko is trying to develop a single strain sensor that can effectively decode an individual's hand gestures.

Professor Sungho Jo reiterated that they could decouple multiple different behaviors observed by a single sensor if they can properly utilize the patterns of human complex motions using machine learning technology.
"After the close collaboration, we were able to develop a single deep-learned sensor that can predict complex hand motions," he added.

The e-skin system developed by the researchers showed highly promising results in the initial evaluations of the study. Complex finger motions were successfully decoded and detected in real-time while consistently operating despite its position on a user's wrist. It was reiterated that the innovation could play an important role in the development of wearable devices such as fitness trackers and the development of robots. 


ⓒ 2024 TECHTIMES.com All rights reserved. Do not reproduce without permission.
Join the Discussion