神经形态工程学
记忆电阻器
计算机科学
数码产品
噪音(视频)
制作
计算
电子工程
电气工程
人工智能
工程类
人工神经网络
图像(数学)
医学
替代医学
病理
算法
作者
Bowen Chen,Jiaping Yao,Jian Guo Xia,Rui Yang,Xiangshui Miao
标识
DOI:10.1109/led.2023.3249340
摘要
Conventional flexible electronics suffer from complex fabrication processes and great energy consumption due to the separated tactile sensors, storage devices, and processing units. In-sensor computing based on multifunctional devices allows efficient computation by directly sensing and processing signals and is considered an effective way to achieve text large-scale integration and low-power devices in the future. Herein, a strain-sensitive multifunctional memristor is prepared based on MoTe2. The device shows good retention (>105 s) and switching stability (>500 cycles). The neuromorphic computing capability of the device was verified using a gesture recognition task with accuracy reaches above 70% even at 50% background noise. The present results demonstrate the attractive multifunction integrated behaviors of MoTe2 devices, opening a new avenue for designing artificially intelligent electronic skins.
科研通智能强力驱动
Strongly Powered by AbleSci AI