Brooks, D;
Sartori, J;
(2017)
Ultra-Low-Power Processors.
IEEE Micro
, 37
(6)
pp. 16-19.
10.1109/MM.2017.4241352.
Preview |
Text
08119709.pdf - Published Version Download (714kB) | Preview |
Abstract
Society's increasing use of connected sensing and wearable computing has created robust demand for ultra-low-power (ULP) edge computing devices and associated system-on-chip (SoC) architectures. In fact, the ubiquity of ULP processing has already made such embedded devices the highest-volume processor part in production, with an even greater dominance expected in the near future. The Internet of Everything calls for an embedded processor in every object, necessitating billions or trillions of processors. At the same time, the explosion of data generated from these devices, in conjunction with the traditional model of using cloud-based services to process the data, will place tremendous demands on limited wireless spectrum and energy-hungry wireless networks. Smart, ULP edge devices are the only viable option that can meet these demands.
| Type: | Article |
|---|---|
| Title: | Ultra-Low-Power Processors |
| Open access status: | An open access version is available from UCL Discovery |
| DOI: | 10.1109/MM.2017.4241352 |
| Publisher version: | https://doi.org/10.1109/MM.2017.4241352 |
| Language: | English |
| Additional information: | Copyright © 2017, IEEE. This is an Open Access article published under the IEEE Open Access Publishing Agreement |
| Keywords: | beyond-CMOS, cognitive computing, convolutional neural network, edge computing device, flying IoT, Internet of Things, security, stochastic computing, system on chip, ultra-low power, visual IoT |
| URI: | https://discovery-pp.ucl.ac.uk/id/eprint/10060248 |
Archive Staff Only
 |
View Item |
