Professor Lee Nae-eung of the Material Engineering, Development of Human Mimicking Artificial Tactors
- 공과대학
- Hit5190
- 2020-06-26
Professor Lee Nae-eung of the Department of New Material Engineering,
Development of Human Mimicking Artificial Tactors with Embedded Intelligence
- Development of intelligent embedded artificial tactile devices by imitating similar synapses of human sensory organs for the first time
- Providing new paradigms for next-generation Neuromorphic sensory recognition systems, intelligent electronic skin, and edge artificial intelligence
Professor Lee Nae-eung's research team (first author Lee Yu-rim, master and master integrated course student of the New Material Engineering Department) announced that it has developed a flexible artificial tactile device that has inherent intelligence by applying a flexible dielectric nanocompound to mimic similar synaptic functions and structures of human sensory organs.
As machine learning and artificial intelligence have recently emerged as key industrial technology areas, research is actively underway to develop sensors that mimic synaptic devices and sensory organs that mimic the human brain, but it is difficult to resolve the bottleneck of fundamental sensor signal processing data as it independently develops information processing processors including synapses and sensors that input information.
In response, the researchers tried to solve this problem by mimicking the similar synaptic connection structure between the Merkel sensory receptor and the sensory neuron end of the human tactile sensory system, and noted that not only the brain but also the sensory organs in the human sensory recognition process do primary information processing on their own and high-level information processing through similar synaptic functions.
The researchers implemented a similar synaptic structure using a nano-particle-polymer composite steel-genetically based transistor structure, and at the same time operated an artificial tactile machine using friction electricity generated when touching the fingertips at the same device's sensors.
Furthermore, using the principle of creating an artificial tactile array and the transistor current recorded after touch, i.e. synaptic weighting, it has been confirmed that adaptability and filtering functions for touch stimuli can be implemented.
It also proved that memory functions similar to sensory memory of human sensory organs can be implemented on the device itself by predicting the number and order of touch, and that the device can be operated reliably by developing it to have flexibility as well as being able to preprocess primary signals on the artificial tactile machine itself.
This study is expected to drastically reduce the data load imposed on processors through pre-processing of signals in the sensor itself, providing a new paradigm for related research such as next-generation neuromorphic sensory recognition system, intelligent electronic skin, and edge-AI. It is also expected that studies imitating human sensory organs will present not only a structural imitation, but also a methodology that implements functionality and intelligence, presenting the direction of future studies on the high-level information processing of neuromorphic sensory recognition systems.
This research was conducted with support from the Ministry of Education's Basic Research Foundation Project (supporting the Center for Research), the Ministry of Science and ICT's Mid-term Research (Type 2 Mid-term Research), and Samsung Electronics (Samsung Strategic Industry-Academic Research), and the results of the research were published in the international journal Nature Communications 6.2 (ed.)
※ Name of the thesis: A flexible automatic-synaptic tactile sensor organ