Search

Your search keyword '"Koji Katayama"' showing total 2 results
Start Over Author "Koji Katayama" Journal 2018 ieee symposium on vlsi technology
2 results on '"Koji Katayama"'

1. From Memory to Sensor: ultra-Low Power and High Selectivity Hydrogen Sensor Based on ReRAM Technology

Author

Satoru Ito, Hisashi Shima, Shinichi Yoneda, Zhiqiang Wei, Hiroyuki Akinaga, Ken Kawai, Homma Kazunari, Satoru Fujii, Koji Katayama, and Yasuhisa Naitoh

Subjects
Materials science, Hydrogen, chemistry.chemical_element, 02 engineering and technology, 01 natural sciences, Hydrogen sensor, chemistry.chemical_compound, Stack (abstract data type), 0103 physical sciences, Electrical and Electronic Engineering, 010302 applied physics, Resistive touchscreen, business.industry, Response time, Disiloxane, 021001 nanoscience & nanotechnology, Electronic, Optical and Magnetic Materials, Resistive random-access memory, Resist, chemistry, Electrode, Optoelectronics, 0210 nano-technology, business, Selectivity, Wireless sensor network
Abstract

We have fabricated a novel hydrogen sensor using optimized 0.18-μm ReRAM process. Our ReHsensor (Resistive Hydrogen Sensor) conforms with the ISO26142 standard in that it exhibits exceptional sensing capabilities, including high sensitivity, wide hydrogen concentration range (up to 4 vol.%) in air and N 2 ambient, high gas selectivity (no reaction with CH 4 , CO, CO 2 , CH 3 OH, and CH 3 COCH 3 ) and is immune to poisoning by SO 2 and hexamethyl disiloxane (HMDS). As it does not require a heater, the power consumption of the ReHsensor is very low, at 0.35 mW. We used this hydrogen sensor device to develop a battery-powered all-in-one wireless hydrogen sensor unit for IoT applications.

Published
2018
Full Text
View/download PDF

2. A 4M Synapses integrated Analog ReRAM based 66.5 TOPS/W Neural-Network Processor with Cell Current Controlled Writing and Flexible Network Architecture

Author

Kazuyuki Kouno, Mochida Reiji, Ono Takashi, Ryutaro Yasuhara, Takumi Mikawa, Hitoshi Suwa, Yasushi Gohou, Koji Katayama, Masayoshi Nakayama, and Hayata Yuriko

Subjects
010302 applied physics, Very-large-scale integration, Network architecture, Artificial neural network, business.industry, Computer science, 020208 electrical & electronic engineering, 02 engineering and technology, Chip, Perceptron, 01 natural sciences, Resistive random-access memory, 0103 physical sciences, 0202 electrical engineering, electronic engineering, information engineering, business, Electrical efficiency, Computer hardware, MNIST database
Abstract

This paper presents low-power neural-network (NN) processor using ReRAM to store weights as analog resistance for future AI computing. We propose ReRAM perceptron circuit for realizing large scale integration, highly accurate cell current controlled writing scheme, and flexible network architecture (FNA) in which any NNs can be configured. Fabricated 180nm test chip shows well-controlled analog cell current with linear 30μA dynamic range and 0.59μA variation of 1 sigma, results in 90.8% MNIST numerical recognition rate. Furthermore, 4M synapses integrated 40nm test chip achieves lower analog cell current and 66.5 TOPS/W power efficiency.

Published
2018
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results