[2PS-253]
Artificially Intelligent Ferroelectric e-Skin for Tactile Learning
발표자신정원 (연세대학교 산학협력단)
연구책임자박철민 (연세대학교)
Abstract
Lightweight, adaptable tactile learning machines are capable of simultaneous sensing, storing, and learning from physical stimuli detected through the skin. These technologies are gaining significant attention for their applications in next-generation wearable and human-interactive neuromorphic electronics. This study presents an integrated AI-powered tactile learning electronic skin, constructed from an array of FeFETs equipped with dome-shaped tactile top gate. When pressure is applied to the gate, the device records changes in ferroelectric remnant polarization within the gate insulator. This leads to analog conductance modulation that varies with both the intensity and frequency of the pressure pulses, effectively emulating various stimuli. The device demonstrates outstanding stability during long-term potentiation and depression cycles, maintaining performance over 10,000 consecutive pulses. Also, low variability of 3.18% ensures consistent and reliable performance.