A Context-aware adaptive algorithm for ambient intelligence DASH at mobile edge computing

Adaptive streaming has recently emerged as a technology enabling high-quality streaming at various bitrates. One of the video streaming challenges remains in research topic nowadays that is choosing optimal segment base on network characteristics and streaming devices, such as network bandwidth, lat...

Full description

Saved in:
Bibliographic Details
Published in:Journal of ambient intelligence and humanized computing 2020-04, Vol.11 (4), p.1377-1385
Main Authors: Kim, Jinsul, Won, Yonggwan, Yoon, Changwoo, Kim, Jin-Young, Park, Sangho, Ryou, JaeCheol, Van Ma, Linh
Format: Article
Language:English
Subjects:
Adaptation
Adaptive algorithms
Ambient intelligence
Artificial Intelligence
Bandwidths
Computational Intelligence
Context
Control algorithms
Edge computing
Engineering
Machine learning
Mobile computing
Multilayer perceptrons
Network latency
Original Research
Quality of service
Robotics and Automation
Segments
Simulation
Streaming media
User Interfaces and Human Computer Interaction
Video transmission
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Adaptive streaming has recently emerged as a technology enabling high-quality streaming at various bitrates. One of the video streaming challenges remains in research topic nowadays that is choosing optimal segment base on network characteristics and streaming devices, such as network bandwidth, latency, the computational capacities of devices. Researchers have proposed many algorithms to overcome such issues within their predefined conditions. However, those proposed methods do not perform efficiently in the heterogeneous network today. Consequently, in this article, we present research on a context-aware adaptive algorithm for ambient intelligence dynamic adaptive employing mobile edge computing (MEC). Specifically, we apply deep learning in the adaptive algorithm which is installed at the MEC to assist clients in choosing the optimal streaming segments as well as reduce network latency. Furthermore, we apply the multilayer perceptron classifier with data obtained from various experiments of adaptive streaming algorithms then combine them in a general algorithm. In the analysis, we use network simulator NS3 as a tool to carry out the verification of our proposed method. As a result, the proposed research reduces network latency as well as improve quality streaming compared to existing approaches.
ISSN:1868-5137
1868-5145
DOI:10.1007/s12652-018-1049-z