Development of electronic skin with AI for augmented human-machine interactions
Newly developed “Substrate-less” design enhanced the performance and wearability of electronic skin
Meta-learning technology allows the electronic skin to recognize various hand gestures regardless of its user
Professor Seung Hwan Ko in the Department of Mechanical Engineering has made a major breakthrough in the development of electronic skin. By integrating biocompatible nanomesh and meta-learning technology, he has created a substrate-less nanomesh receptor capable of rapidly recognizing various human hand motions, regardless of the user. The findings are expected to revolutionize the field of wearable electronics, particularly in healthcare engineering, such as rehabilitation, human augmentation, and robotics.
The human hand is an essential part of the body that is extensively used in performing daily tasks such as grasping objects, manipulating tools, and carrying out other activities of daily living. Any injury or restriction in the movement of the hand can significantly impact the individual’s ability to perform these tasks and affect their quality of life. Therefore, rapid and precise recognition of hand gestures plays a critical role in healthcare engineering, aiding the rehabilitation of patients with physical disabilities and enabling human augmentation in an aging society.
Wearable mechanical sensors have provided suitable human-to-machine platforms to recognize human hand-based motions by converting mechanical stimuli into electrical signals. However, accurately classifying signals generated by random human hand-based movements can be challenging due to the high degree of freedom of hand motion and the vast amount of random information detected by the sensor. Recent developments in AI technologies have assisted in classifying this enormous amount of data, demonstrating the potential to recognize hand gestures and tasks with wearable electronics such as electromyography wristbands or electronic gloves. These developments in advanced AI technology in wearble electronic devices are heading towards seamless integration into practical healthcare applications.
Prof. Ko has developed an innovative epidermal sensor by directly printing biocompatible nanomesh onto human skin. This eliminated the need for a supporting substrate previously inevitable in traditional wearable sensor designs. The sensor's substrate-less design provides a comfortable user environment with excellent mechanical compliance, making it imperceptible to users who are using the technology. The sensor can detect even the subtlest hand movements and motions associated with human hand gestures, with electrical signals wirelessly transmitted to the AI for further data processing.
In addition, the meta-learning algorithm, a cutting-edge AI technology, was adapted to train the electronic skin with signals from hand gestures. The meta-learning technology enables the identification of unlabeled signals and the detection of unique signal patterns from random hand gestures. This approach enabled the electronic skin to learn from minimal training, significantly increasing the accuracy and speed of learning. The developed system has demonstrated practical applications of real-time hand motion tracking, such as a virtual keyboard and recognition of random objects with high accuracy and speed. Due to the meta-learning technology, the system minimized the amount of data collection for new users.
The developed seamless human-to-machine system with AI-equipped biocompatible nanomesh is expected to pave the way for advanced healthcare applications such as remote physical therapy and physician training, as it can be applied to any part of the human body. Furthermore, combined with a machine-to-human system such as wearable soft robotics, these findings can provide more powerful applications, including telemedicine which requires reciprocal interaction between humans and machines.
Prof. Ko and his team are currently researching various wearable applications with AI technology for advanced human-machine interactions. Their developments hold great promise in accelerating progress in the field of wearable electronics, including healthcare applications and robotics. This brings us one step closer to a future where human-machine interactions are seamless and intuitive in our daily life.
The research findings were published in January in <Nature Electronics>, the world’s most prestigious academic journal in the field of electronics.
For further information, please contact Prof. Seung Hwan Ko (email@example.com).
Overall view of hand gesture recognition with seamless integration of
biocompatible nanomesh and meta-learning technology and
its practical applications with advanced human-machine interfaces
1. K.K. Kim and M. Kim et al. Nat. Electron. 2023. 1-12.
2. K.K. Kim et al. Adv. Intell. Syst. 2020, 3, 200157
3. K.K. Kim, IH Ha, M. Kim et al. Nat. Commun. 2020. 11, 2149