神经形态工程学
材料科学
光电子学
氧化物
计算机科学
人工智能
人工神经网络
冶金
作者
Qihan Liu,Li Yin,Chun Zhao,Jingyi Wang,Ziang Wu,Lei Hao,Yina Liu,Bowen Tian,Zhiyuan Zhang,Zishen Zhao,Ruofu Liu,Changzeng Ding,Yunfei Han,Chang‐Qi Ma,Pengfei Song,Ivona Z. Mitrović,Eng Gee Lim,Zhen Wen
出处
期刊:Nano Energy
[Elsevier BV]
日期:2022-08-13
卷期号:102: 107686-107686
被引量:46
标识
DOI:10.1016/j.nanoen.2022.107686
摘要
Human central nervous system and the peripheral nervous system have played significant roles in mediating the interactions with the outside world. Inspired by the human nervous systems, artificial sensory and neuromorphic innovations have been developed to mimic nervous functions. A hybrid mixed-dimensional perovskite/metal-oxide heterojunction has been demonstrated in this work for three-terminal all-in-one opto-electric artificial synapse to integrate opto-electric synaptic emulations and optical perception functions. Based on the well-designed layer configuration, an all-in-one device, consisting of the ion-electrolyte layer, ion-permeable metal-oxide semiconductor channel layer, the mixed-dimensional perovskite optical perception layer, and the amorphous ZnO passivation layer, has been demonstrated with superior electrical performance. Utilizing an ion-electrolyte and ion-permeable metal-oxide semiconductor structure, the synaptic emulation modulated by electrical gate-stimulus could be effectively achieved. The optical perception and synaptic plasticity modulated by the optical stimulus have been integrated into the all-in-one device. Furthermore, the electrolyte gated device enables artificial visual adaptation with adaptive behavior of environmental lightness under dim and bright conditions. Addition, artificial visual persistence has been emulated by the device utilizing the optical synaptic behavior. Based on those properties, a cascaded near-sensor face recognition access control retinal-neuromorphic computing system has been developed based on the all-in-one device. The retinal-neuromorphic system based on all-in-one devices could recognize the face of a requester with an accuracy rate over 90 %, and ignore the passers with the lingering phenomenal trace as the result of the artificial visible persistence.
科研通智能强力驱动
Strongly Powered by AbleSci AI