|
Title:
|
APPLY EDGE INTELLIGENCE TO IOT BASED HOME
AUTOMATION |
|
Author(s):
|
Yi Hua Wu and Hongli Luo |
|
ISBN:
|
978-989-8704-24-5 |
|
Editors:
|
Hans Weghorn |
|
Year:
|
2020 |
|
Edition:
|
Single |
|
Keywords:
|
Internet of Things, Edge Computing, Deep Learning, Video Surveillance, Privacy |
|
Type:
|
Short |
|
First Page:
|
129 |
|
Last Page:
|
133 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
There are many challenges in building a low-delay IoT system because the reliance on underlying network infrastructure
and data analytics in cloud environment. Deep learning combined with edge computing makes edge intelligence possible
in IoT system. With deep learning, complicated task including facial recognition and voice control can be processed at
edge devices such as mobile phone and microcontroller. In this paper, we present a design and implementation of home
automation system which applies deep learning techniques at low-power IoT edge devices. We implemented the facial
recognition in TensorFlow Lite on Raspberry Pi. The performance of TensorFlow and TensorFlow Lite on Raspberry Pi
is also compared. The experimental data we collected from the system proves that the performance can be efficiently
improved using deep learning models in TensorFlow Lite. With all data constrained at local area network, it can reduce
processing delay and computation resource consumption, while at the same time provide real-time response and better
privacy. |
|
|
|
|
|
|
