Machine Learning and Intelligent Communications Part II 2017 pdf pdf
Xuemai Gu Gongliang Liu (Eds.)
Bo Li 227
Machine Learning and Intelligent Communications Second International Conference, MLICOM 2017 Weihai, China, August 5–6, 2017 Proceedings, Part II
Lecture Notes of the Institute for Computer Sciences, Social Informatics
and Telecommunications Engineering 227
Editorial BoardOzgur Akan Middle East Technical University, Ankara, Turkey
Paolo Bellavista University of Bologna, Bologna, Italy
Jiannong Cao Hong Kong Polytechnic University, Hong Kong, Hong Kong
Geoffrey Coulson Lancaster University, Lancaster, UK
Falko Dressler University of Erlangen, Erlangen, Germany
Domenico Ferrari Università Cattolica Piacenza, Piacenza, Italy
Mario Gerla UCLA, Los Angeles, USA
Hisashi Kobayashi Princeton University, Princeton, USA
Sergio Palazzo University of Catania, Catania, Italy
Sartaj Sahni University of Florida, Florida, USA
Xuemin Sherman Shen University of Waterloo, Waterloo, Canada
Mircea Stan University of Virginia, Charlottesville, USA
Jia Xiaohua City University of Hong Kong, Kowloon, Hong Kong
Albert Y. Zomaya University of Sydney, Sydney, Australia More information about this series at http://www.springer.com/series/8197
Xuemai Gu Gongliang Liu Bo Li (Eds.) Machine Learning and Intelligent Communications
Second International Conference, MLICOM 2017 Weihai, China, August 5–6, 2017 Proceedings, Part II Editors Xuemai Gu Bo Li Harbin Institute of Technology Shandong University Harbin, Heilongjiang Weihai, Heilongjiang China China Gongliang Liu Harbin Institute of Technology Weihai, Heilongjiang China
ISSN 1867-8211
ISSN 1867-822X (electronic) Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering
ISBN 978-3-319-73446-0
ISBN 978-3-319-73447-7 (eBook) https://doi.org/10.1007/978-3-319-73447-7 Library of Congress Control Number: 2017963764 ©
ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2018
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the
material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now
known or hereafter developed.The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.The publisher, the authors and the editors are safe to assume that the advice and information in this book are
believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors
give a warranty, express or implied, with respect to the material contained herein or for any errors or
omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations. Printed on acid-free paper
Preface
We are delighted to introduce the proceedings of the second edition of the 2017 European Alliance for Innovation (EAI) International Conference on Machine Learning and Intelligent Communications (MLICOM). This conference brought together researchers, developers, and practitioners from around the world who are leveraging and developing machine learning and intelligent communications.
The technical program of MLICOM 2017 consisted of 141 full papers in oral presentation sessions at the main conference tracks. The conference tracks were: Main Track, Machine Learning; Track 1, Intelligent Positioning and Navigation; Track 2, Intelligent Multimedia Processing and Security; Track 3, Intelligent Wireless Mobile Network and Security; Track 4, Cognitive Radio and Intelligent Networking; Track 5, Intelligent Internet of Things; Track 6, Intelligent Satellite Communications and Net- working; Track 7, Intelligent Remote Sensing, Visual Computing and Three-Dimensional Modeling; Track 8, Green Communication and Intelligent Net- working; Track 9, Intelligent Ad-Hoc and Sensor Networks; Track 10, Intelligent Resource Allocation in Wireless and Cloud Networks; Track 11, Intelligent Signal Processing in Wireless and Optical Communications; Track 12, Intelligent Radar Signal Processing; Track 13, Intelligent Cooperative Communications and Networking. Aside from the high-quality technical paper presentations, the technical program also featured three keynote speeches. The three keynote speeches were by Prof. Haijun Zhang from the University of Science and Technology Beijing, China, Prof. Yong Wang from Harbin Institute of Technology, China, and Mr. Lifan Liu from National Instruments China.
Coordination with the steering chairs, Imrich Chlamtac, Xuemai Gu, and Gongliang Liu, was essential for the success of the conference. We sincerely appreciate their constant support and guidance. It was also a great pleasure to work with such an excellent Organizing Committee who worked hard to organize and support the con- ference, and in particular, the Technical Program Committee, led by our TPC co-chairs, Prof. Xin Liu and Prof. Mingjian Sun, who completed the peer-review process of technical papers and created a high-quality technical program. We are also grateful to the conference manager, Katarina Antalova, for her support and to all the authors who submitted their papers to MLICOM 2017.
We strongly believe that the MLICOM conference provides a good forum for researchers, developers, and practitioners to discuss all the science and technology aspects that are relevant to machine learning and intelligent communications. We also hope that future MLICOM conferences will be as successful and stimulating, as indicated by the contributions presented in this volume. December 2017
Xuemai Gu
Organization
Steering CommitteeSteering Committee Chair Imrich Chlamtac University of Trento, Create-Net, Italy Steering Committee Xin-Lin Huang Tongji University, China
Organizing Committee
General Chairs Xuemai Gu Harbin Institute of Technology, China Z. Jane Wang The University of British Columbia, Canada Gongliang Liu Harbin Institute of Technology (Weihai), China General Co-chairs Jianjiang Zhou Nanjing University of Aeronautics and Astronautics,
China Xin Liu Dalian University of Technology, China Web Chairs Xuesong Ding Harbin Institute of Technology (Weihai), China Zhiyong Liu Harbin Institute of Technology (Weihai), China Xiaozhen Yan Harbin Institute of Technology (Weihai), China Publicity and Social Media Chair Aijun Liu Harbin Institute of Technology (Weihai), China Sponsorship and Exhibits Chair Chenxu Wang Harbin Institute of Technology (Weihai), China Publications Chairs Xin Liu Dalian University of Technology, China Bo Li Harbin Institute of Technology (Weihai), China Posters and PhD Track Chair Local Chair Bo Li Harbin Institute of Technology (Weihai), China Conference Manager Katarina Antalova EAI - European Alliance for Innovation
Technical Program Committee
Technical Program Committee Chairs Z. Jane Wang University of British Columbia, Canada Xin Liu Dalian University of Technology, China Mingjian Sun Harbin Institute of Technology (Weihai), China TPC Track Chairs Machine Learning Xinlin Huang Tongji University, China Rui Wang Tongji University, China Intelligent Positioning and Navigation Mu Zhou Chongqing University of Posts and Telecommunications, China Zhian Deng Dalian Maritime University, China Min Jia Harbin Institute of Technology, China Intelligent Multimedia Processing and Security Bo Wang Dalian University of Technology, China Fangjun Huang Sun Yat-Sen University, China Wireless Mobile Network and Security Shijun Lin Xiamen University, China Yong Li Tsinghua University, China Cognitive Radio and Intelligent Networking Yulong Gao Harbin Institute of Technology, China Weidang Lu Zhejiang University of Technology, China Huiming Wang
Xi’an Jiaotong University, China Intelligent Internet of Things Xiangping Zhai Nanjing University of Aeronautics and Astronautics,
China Chunsheng Zhu The University of British Columbia, Canada Yongliang Sun Nanjing Tech University, China Intelligent Satellite Communications and Networking Kanglian Zhao Nanjing University, China Zhiqiang Li PLA University of Science and Technology, China
VIII Organization
Organization
IX
Intelligent Remote Sensing, Visual Computing, and Three-Dimensional Modeling Jiancheng Luo Institute of Remote Sensing and Digital Earth,
Chinese Academy of Sciences, China Bo Wang Nanjing University of Aeronautics and Astronautics,
China Green Communication and Intelligent Networking Jingjing Wang Qingdao University of Science and Technology, China Nan Zhao Dalian University of Technology, China Intelligent Ad-Hoc and Sensor Networks Bao Peng Shenzhen Institute of Information Technology, China Danyang Qin Heilongjiang University, China Zhenyu Na Dalian Maritime University, China Intelligent Resource Allocation in Wireless and Cloud Networks Feng Li Zhejiang University of Technology, China Jiamei Chen Shenyang Aerospace University, China Peng Li Dalian Polytechnic University, China Intelligent Signal Processing in Wireless and Optical Communications Wei Xu Southeast University, China Enxiao Liu Institute of Oceanographic Instrumentation,
Shandong Academy of Sciences, China Guanghua Zhang Northeast Petroleum University, China Jun Yao Broadcom Ltd., USA Intelligent Radar Signal Processing Weijie Xia Nanjing University of Aeronautics and Astronautics,
China Xiaolong Chen Naval Aeronautical and Astronautical University,
China Intelligent Cooperative Communications and Networking Deli Qiao East China Normal University, China Jiancun Fan
Xi’an Jiaotong University, China Lei Zhang University of Surrey, UK
Contents – Part II
Intelligent Resource Allocation in Wireless and Cloud Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Daxing Qian, Ximing Pei, and Xiangkun Li . . . . . . . . . . . . . . . . . . . . . .
Jianfei Shi, Feng Li, Xin Liu, Mu Zhou, Jiangxin Zhang, and Lele Cheng . . . .
Li Wang, Lele Cheng, Feng Li, Xin Liu, and Di Shen . . . . . . . . . .
Xujie Li, Xing Chen, Ying Sun, Ziya Wang, Chenming Li, and Siyang Hua . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . .
Dawei Chen, Enwei Xu, Shuo Shi, and Xuemai Gu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Xin-yong Yu, Ying Guo, Kun-feng Zhang, Lei Li, Hong-guang Li, and Ping Sui . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Zijian Zhang, Dongxuan He, and Yulei Nie . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Yi Wang, Xiangyuan Bu, Xiaozheng Gao, and Lu Tian XII Contents – Part II
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . .
Jiangang Wen, Jingyu Hua, Zhijiang Xu, Weidang Lu, and Jiamin Li
Yulei Nie, Zijian Zhang, and Peipei Liu Intelligent Radar Signal Processing
Yuxue Sun, Ying Luo, and Song Zhang
Han-yang Xu and Feng Zhou
Qi-fang He, Han-yang Xu, Qun Zhang, and Yi-jun Chen
Feng Zhu, Xiaofeng Hu, Xiaoyuan He, Kaiming Li, and Lu Yang
Jia-cheng Ni, Qun Zhang, Li Sun, and Xian-jiao Liang
Fu-fei Gu, Le Kang, Jiang Zhao, Yin Zhang, and Qun Zhang
Di Meng, Han-yang Xu, Qun Zhang, and Yi-jun Chen
Jie Pan, Yalin Zhu, and Changling Zhou
Jian Luo, Honggang Zhang, and Yuanyuan Song
XIII Contents – Part II
Yi-shuai Gong, Qun Zhang, Kai-ming Li, and Yi-jun Chen
Luyao Cui, Aijun Liu, Changjun Yv, and Taifan Quan
Ying Zhou, Weijie Xia, Jianjiang Zhou, Linlin Huang, and Minling Huang
Xiaolong Chen, Xiaohan Yu, Jian Guan, and You He
Yang Xuguang, Yu Changjun, Liu Aijun, and Wang Linwei Intelligent Cooperative Communications and Networking
Fangni Chen, Jingyu Hua, Weidang Lu, and Zhongpeng Wang
Yuan Feng, Fu-sheng Dai, and Ji Zhou
Zongjie Bi, Zhaoshuo Tian, Pushuai Shi, and Shiyou Fu
Hong Peng, Changran Su, Yu Zhang, Linjie Xie, and Weidang Lu The Second Round
Bei Cao and Yongsheng Wang
Yao Zhang, Chenxu Wang, Xinsheng Wang, Jing Wang, and Le Man
Chenguang He, Yuwei Cui, and Shouming Wei
Dongxing Bao, Xiaoming Li, and Jin Li
Dongxing Bao, Xiaoming Li, Yizong Xin, Jiuru Yang, Xiangshi Ren, Fangfa Fu, and Cheng Liu
Yanguo Zhou, Hailin Zhang, Ruirui Chen, and Tao Zhou
Shuai Shao, Changjun Yu, and Kongrui Zhao
Xiaojuan Guo and Xiyu Liu
Yu Xu, Dezhi Li, Zhenyong Wang, Gongliang Liu, and Haibo Lv
Xuanxuan Tian, Tingting Zhang, Qinyu Zhang, Hongguang Xu, and Zhaohui Song
Hui Li, Zhigang Gai, Enxiao Liu, Shousheng Liu, Yingying Gai, Lin Cao, and Heng Li
Xin Zhang and Hang Dong
Tong Xue and Yong Liu
XIV Contents – Part II
XV Contents – Part II
Baobao Wang, Haijun Zhang, Keping Long, Gongliang Liu, and Xuebin Li
Xinwu Chen, Yaqin Xie, Erfu Wang, and Danyang Qin
Haicheng Qu, Jitao Qin, Wanjun Liu, and Hao Chen
Dongqing Li, Junxin Luo, Tiantian Zhang, Shaohua Wu, and Qinyu Zhang
Xiao Luo, Xinhong Wang, Ping Wang, Fuqiang Liu, and Nguyen Ngoc Van
Weizhi Zhong, Lei Xu, Xiaoyi Lu, and Lei Wang
Weihao Xie, Zhigang Gai, Enxiao Liu, and Dingfeng Yu
Haowei Li, Liming Zheng, Yue Wu, and Gang Wang
Ligang Cong, Huamin Yang, Yanghui Wang, and Xiaoqiang Di
MingFeng Wang, AiJun Liu, LinWei Wang, and ChangJun Yu
Runxuan Li, Yu Sun, and Qingquan Sun XVI Contents – Part II
Yongliang Sun, Yejun Sun, and Kanglian Zhao
Yongsheng Wang, Chen Yin, and Xunzhi Zhou
Jiaxin Chen, Yuhua Xu, Yuli Zhang, and Qihui Wu
Liyong Ma, Lidan Tang, Wei Xie, and Shuhao Cai
Naizhang Feng, Teng Jiang, Shiqi Duan, and Mingjian Sun
Mo Han, Jun Shi, Yiqiu Deng, and Weibin Song
Bing Zhao and Ganlin Hao
Bing Zhao
Yu Zhang, Yangyang Li, Ruide Li, and Wenjing Sun
Xinwu Chen, Yaqin Xie, and Erfu Wang
Ruofei Zhou, Gang Wang, Wenchao Yang, Zhen Li, and Yao Xu
Fangfang Cheng, Jiyu Jin, Guiyue Jin, Peng Li, and Jun Mou
XVII Contents – Part II
Sichen Zhao, Yuan Fang, Wenfeng Li, and Kanglian Zhao
Feng Qi and Mengmeng Liu Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Contents – Part I
Machine Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Jiamei Chen, Yao Wang, Xuan Li, and Chao Gao . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Qingquan Sun, Jiang Lu, Yu Sun, Haiyan Qiao, and Yunfei Hou . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Hang Dong and Xin Zhang . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Xinyou Li, Wenjing Kang, and Gongliang Liu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Hongxu Zheng, Jianlun Wang, and Can He . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Liming Zheng, Donglai Zhao, Gang Wang, Yao Xu, and Yue Wu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Yao Xu, Gang Wang, Liming Zheng, Rongkuan Liu, and Donglai Zhao Intelligent Positioning and Navigation . . . . . . .
XX Contents – Part I
Sunan Li, Jingyu Hua, Feng Li, Fangni Chen, and Jiamin Li
Zengshan Tian, Yong Li, Mu Zhou, and Yinghui Lian
Mingjian Sun, Shengmiao Lv, Xue Zhao, Ruya Li, Wenhan Zhang, and Xiao Zhang
Mu Zhou, Xiaoxiao Jin, Zengshan Tian, Haifeng Cong, and Haoliang Ren Intelligent Multimedia Processing and Security
Ge Liu, Fangjun Huang, Qi Chen, and Zhonghua Li
Fulong Yang, Yabin Li, Kun Chong, and Bo Wang
Yabin Li, Bo Wang, Kun Chong, and Yanqing Guo Wireless Mobile Network and Security
Chungang Liu, Chen Wang, and Wenbin Zhang
Zhe Li, Yanxin Yin, and Lili Wu Cognitive Radio and Intelligent Networking
Yunxue Gao, Liming Zheng, Donglai Zhao, Yue Wu, and Gang Wang
XXI Contents – Part I
Jingting Xiao, Ruoyu Zhang, and Honglin Zhao
Fei An and Fusheng Dai
Yanping Chen, Yulong Gao, and Yongkui Ma
Jingming Li, Guoru Ding, Xiaofei Zhang, and Qihui Wu
Hui Kang, Hongyang Xia, and Fugang Liu Intelligent Internet of Things
Ershi Xu, Xiangping Zhai, Weiyi Lin, and Bing Chen
Hongyuan Wang
Hang Wang, Yu Sun, and Qingquan Sun
Yongan Guo, Tong Liu, Xiaomin Guo, and Ye Yang
Mingze Xia and Dongyu Song
Qi Zhang, Zhiqiu Huang, and Jian Xie Intelligent Satellite Communications and Networking
Xiaoqin Ni, Kanglian Zhao, and Wenfeng Li XXII Contents – Part I
Dongxu Hou, Kanglian Zhao, and Wenfeng Li
Xiaolin Xu, Yu Zhang, and Jihua Lu
Beishan Wang and Qi Guo
Yu Xu, Dezhi Li, Zhenyong Wang, Gongliang Liu, and Haibo Lv
Bo Li, Xiyuan Peng, Hongjuan Yang, and Gongliang Liu
Yumeng Zhang, Mingchuan Yang, and Xiaofeng Liu
Wenrui Zhang, Chenyang Fan, Kanglian Zhao, and Wenfeng Li Intelligent Remote Sensing, Visual Computing and Three-Dimensional Modeling
Yihao Wang, Yuncui Zhang, Xufen Xie, and Yuxuan Zhang
Xu Huang, Yanfeng Zhang, Gang Zhou, Lu Liu, and Gangshan Cai
Yilang Sun, Shuqiao Sun, Zihao Cui, Yanchao Zhang, and Zhaoshuo Tian
Hongda Fan, Xufen Xie, Yuncui Zhang, and Nianyu Zou
XXIII Contents – Part I
Green Communication and Intelligent Networking
Changjun Chen, Jianxin Dai, Chonghu Cheng, and Zhiliang Huang
Kaijian Li, Jianxin Dai, Chonghu Cheng, and Zhiliang Huang
Xudong Yin, Jianxin Dai, Chonghu Cheng, and Zhiliang Huang
Juan Liu, Jianxin Dai, Chonghu Cheng, and Zhiliang Huang
Mingze Xia
Min Zhang, Jianxin Dai, Chonghu Cheng, and Zhiliang Huang
Xinyu Zhang, Jing Guo, Qiuyi Cao, and Nan Zhao
Xin Liu, Xiaotong Li, Zhenyu Na, and Qiuyi Cao Intelligent Ad-Hoc and Sensor Networks
Zhuangguang Chen and Bei Cao
Tong Liu, Zhimou Xia, Shuo Shi, and Xuemai Gu
Bei Cao, Tianliang Xu, and Pengfei Wu
Shuang Wu, Zhenyong Wang, Dezhi Li, Qing Guo, XXIV Contents – Part I
Shuang Wu, Zhenyong Wang, Dezhi Li, Gongliang Liu, and Qing Guo
Guoqiang Wang and Bai Sun
Juan Chen, Zhengkui Lin, Xin Liu, Zhian Deng, and Xianzhi Wang
Jiaqi Zhen, Yong Liu, and Yanchao Li
Jiaqi Zhen, Yong Liu, and Yanchao Li
Rui Du, Wenjing Kang, Bo Li, and Gongliang Liu
Danyang Qin, Ping Ji, Songxiang Yang, and Qun Ding
Yang Li and Peidong Zhuang
Weiguang Zhao and Peidong Zhuang
Songyan Liu, Ting Chen, Shangru Wu, and Cheng Zhang
Songyan Liu, Shangru Wu, Ting Chen, and Cheng Zhang
XXV Contents – Part I
Yan Wu, Wenjing Kang, Bo Li, and Gongliang Liu
Zhongchao Ma, Liang Ye, and Xuejun Sha
Danyang Qin, Songxiang Yang, Ping Ji, and Qun Ding
Guoqiang Wang and Shangfu Li Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Intelligent Resource Allocation in Wireless and Cloud Networks
The Application of Equivalent Mean Square
Error Method in Scalable Video Perceptual
Quality
(&)
Daxing Qian , Ximing Pei, and Xiangkun Li
Dalian Neusoft University of Information,
Software Park Road 8, Dalian 116023, Liaoning, China
{qiandaxing,peiximing,lixiangkun}@neusoft.edu.cn
Abstract. Scalable video is a stream video over heterogeneous networks to
different clients. To provide the better quality of service (QoS) or quality of
experience (QoE) to customer, we propose an Equivalent Mean Square Error
(Eq-MSE) method which is developed based on spatial and temporal frequency
analysis of input video content. Eq-MSE is used to calculate minimal frame rate
(MinFR) for different videos to guarantee motion without jitter. The proposed
scheme in this paper can provide better perceptual video quality than without
considering the video content impact. Keywords: SVC Eq-MSE MinFR Perceptual quality1 Introduction
With the advances of semi-conductor and access network technologies, real-time video streaming becomes more and more popular in our daily life. We can enjoy the videos service at famous website through different networks using heterogeneous devices. How to provide the high quality of service (QoS) or quality of experience (QoE) to different users over heterogeneous networks is a crucial problem for the success of video streaming application. Scalable video coding (SVC)
is a full resolution
scalable video stream which can be truncated to adapt different requirements imposed by the subscribed users and underlying access networks.
SVC includes temporal, spatial, SNR and combined scalabilities. Temporal scala- bility is realized by the hierarchical-B prediction [
Spatial scalability is achieved by
encoding each supported spatial resolution into one layer. SNR scalability includes coarse grain scalability (CGS) and medium grain scalability (MGS) [
To achieve the
SNR refinement, we usually use different quantization steps at different SNR layers. In this paper, we study the temporal and SNR joint scalability, and the spatial scalability is not mentioned.
Video content have a significant impact on the perceptual quality. For example, a motion intensive video need a larger frame rate to maintain the continuity of the object movement and avoid jitter and guarantee the motion smoothness, while for stationary
4
D. Qian et al.
other hand, if there are larger high-frequency components (i.e., rich texture) in a single frame of the video, a finer quantization to reach better spatial quality is typically preferred. To solve the problem, we study the spatial and temporal frequency of the input video content, and propose an Equivalent Mean Square Error (Eq-MSE) scheme to derive the minimal frame rate (MinFR) for different video sources to guarantee the motion smoothness and excellent QoE of the decoded video.
This paper is organized as follows. Section
we introduce the Eq-MSE method to derive
the minimal frame rate without jitter for different input video sources. Subjective test evaluation and experimental results are shown in Sect.
concludes the paper and discusses the future directions.
2 Temporal Frequency
The concept of spatial frequency is introduced in [
We can use the function ; f ; f W f x y t ZZZ
¼ w x ð ; y ; t Þ exp j 2p f x þ f y þ f t ZZ x y t dxdydt w x v t ; y v t j 2p f x v t y v t
¼ x y exp x ð x Þ þ f y y dxdy Z ð1Þ exp j 2p f v v Z t þ f x x þ f y y t dt f ; f exp j 2p f v v
x y t þ f x x þ f y y t dt
¼ W f ; f v v
x y d f t þ f x x þ f y y
¼ W where f ; f x ; y W x y indicates the 2D CSFT of w ð Þ. This function means that a spatial pattern characterized by f ; f in the object will lead to a temporal frequency, i.e.,
x y
f t ¼ f x v x f y v y ð2Þ For a video signal, the temporal frequency is 2D position dependent. For a fixed 2D position (x, y), its temporal frequency is defined as the number of cycles per second usually denoted by Hertz (Hz).
From
) we can draw a conclusion that the temporal frequency depends on not
only the motion, but also the spatial frequency [ of the object.
3 Equivalent Mean Square Error (Eq-MSE)
The Application of Eq-MSE Method in Scalable Video Perceptual Quality
5 Fig. 1. Illustrative figure for object in general picture.
Figure illustrates that the size of black column is w h and picture size is W H. We use f tB to represent the induced frame rate by the object, which is defined as:
h w
MSEf ; MSEf ;
H W
f
tB ¼ v x v y ð3Þ
MSEf 1 ; MSEf ;
1 ð Þ ð Þ where MSEf f ; ; y ; ; f is
ð x Þ is MSE x ð Þ when the picture SF is f ð x Þ, and MSEf 0 y MSE x ð ; y Þ when the picture SF is 0 ; f . v and v are velocities in horizontal and
y x y vertical directions.
We regard the SF of a general picture is:
h w
MSEf ; MSEf ; h w
; ; ;
H W f x f y ¼ ¼ ð4ÞMSEf 1 ; MSEf ;
1 H W ð Þ ð Þ
The objects in a picture that induce the frame rate from moving from arbitrary directions are: X X MSEf ; MSEf ;
h w
P H W h w f f v v v vt ¼ tB ¼ x y ¼ x y ð5Þ H W
MSEf 1 ; MSEf ;
1 ð Þ ð Þ where P is all the MBs in the picture. v and v are velocities in horizontal and vertical
x y
directions of corresponding MB. We get the mode and number of MB in a picture, and then choose the other picture within the same GOP to get MVs according to every MB. The ratio between MVs number in MB and the time interval between two frames are v x P P
h w
and v . For example, the and of sequence Mobile are 12.2, 9.4,
y v x v y H W
respectively. Its MinFR is 12.2 + 9.4 = 21.6. Note that with a real signal, the CSFT is symmetric, so that for every frequency component at f ; f
x y , there is also a component
at f ; f with the same magnitude. The corresponding temporal frequency caused
x y
by this other component is f v þ f v [
x x y y
Equation
) is the function of minimal frame rate (MinFR) that makes the video motion smoothness without jitter.
6
D. Qian et al.
We invite 15 experimenters to give the decoded video subjective ratings for evaluate the subjective quality. Sub0 is the default scalable video adaptation without considering the video content impact, while sub1 is scalable adaptation with dependent video content. We use 11 ranks (i.e., 0–10) for the subjective tests ranging. The worst is 0 and the best is 10. The subjective assessment follows
.
Table 1. Sequences subjective test comparative results
Sequences Sub0 Sub1
Akiyo
6.6
6.8 City
4.9
7.7 Mobile
5.7
7.1 Football
6.1
7.9 Table depicts the subjective test results of four sequences. It is obviously that
“City”, “Mobile” and “Football” have better perceptual rating for sub1 session, while “Akiyo” is quite similar between sub1 and sub0. We can draw a conclusion that the Eq-MSE method is providing better-decoded video quality at a given bit rate.
5 Conclusions
In this paper, we propose the Eq-MSE scheme, which is developed based on the spatial and temporal frequency analysis of the video content. This scheme is used to derive the MinFR for different videos and in consequence, so as to guarantee the motion smoothness for decent decoded video quality. Compared with the default scalable video adaptation without considering the video content impact, our proposed scheme can provide better perceptual video quality at the same bit rate according to the sub- jective quality assessments.
References
1. Text of ISO/IEC 14496-10:2005/FDAM 3 Scalable Video Coding, Joint Video Team
(JVT) of ISO-IEC MPEG and ITU-T VCEG, Lausanne, N9197 (2007)
2. ISO/IEC ITU-T Rec. H264: Advanced Video Coding for Generic Audiovisual Services, Joint
Video Team (JVT) of ISO-IEC MPEG and ITU-T VCEG, International Standard (2003)
3. Schwarz, H., Marpe, D., Wiegand, T.: Hierarchical B pictures. In: Joint Video Team, Doc.
JVT-P014 (2005)
4. Schwarz, H., Marpe, D., Wiegand, T.: Overview of the scalable video coding extension of the
H.264/AVC standard. IEEE Trans. Circ. Syst. Video Technol. 17(9), 1103–1120 (2007)
5. Wang, Y., Ostermann, J., Zhang, Y.-Q.: Video Processing and Communications (2001)
6. Qian, D., Wang, H., Niu, F.: Scalable video coding bit stream extraction based on equivalent
MSE method. In: Advanced Materials Research, vol. 204–210, pp. 1728–1732 (2011)
7. Qian, D., Wang, H., Sun, W., Zhu, K.: Bit stream extraction based on video content method in
the scalable extension of H.264/AVC. J. Softw. 6, 2090–2096 (2011)
8. ITU-R Rec. BT.500-11: Methodology for the subjective assessment of the quality of
television pictures (2002)The Application of Eq-MSE Method in Scalable Video Perceptual Quality
7
Spectrum Allocation in Cognitive Radio
Networks by Hybrid Analytic Hierarchy Process
and Graph Coloring Theory
3 Chongqing University of Posts and Telecommunications, Chongqing 400065, China
zhoumu@cqupt.edu.cn
Abstract.proposed an improved graph coloring algorithm for spectrum allocation with
]
as the list coloring algorithm can allocate only one spectrum once, the authors proposed a spectrum allocation algorithm to assign channels to multiple users at the same time without incurring interference. Based on
,
includes distributed greedy algorithm and distributed fairness algorithm. In
] proposed a list-coloring algorithm based on the graph coloring theory, which
]. Graph theory, as a classical optimization theory, has been introduced to solve the difficulty of spectrum allocation in cognitive radio.
Spectrum sharing is the key technology in cognitive radio which attracts the increasing interest
Spectrum allocation Analytic hierarchy process
Graph coloring ·
Keywords: Cognitive radio ·
In this paper, a graph coloring-based spectrum allocation
algorithm in cognitive radio networks combined with analytic hierarchy
process is proposed. By analyzing several key factors that affect the qual-
ity of the leased spectrum, the algorithm combines the graph algorithm
and analytic hierarchy process to assign the optimal spectrum to cogni-
tive users orderly. Simulation results show that the proposed algorithm
can effectively improve the network efficiency compared with original
algorithms and arose inconspicuous loss to the whole network’s fairness.
The proposal not only improves the efficiency of spectrum allocation, but
also balances the requirements of the overall fairness of cognitive radio
networks.2 Dalian University of Technology, Dalian 116024, Liaoning, China
liuxinstar1984@dlut.edu.cn
Jianfei Shi
1 Zhejiang University of Technology, Hangzhou 310023, Zhejiang, China
{sjf,fenglzj,zjx}@zjut.edu.cn, 478708892@qq.com
1
, and Lele Cheng
1
, Jiangxin Zhang
3
, Mu Zhou
2
, Xin Liu
)
1(
B
, Feng Li
1
1 Introduction
Spectrum Allocation in Cognitive Radio Networks by Hybrid AHP
9
In this paper, we apply analytic hierarchy process (AHP), one of the multi- objective decision method to proceed spectrum decision and provide a reasonable spectrum access strategy according to the heterogeneous idle spectrum. The improved proposal takes into account diverse spectral characteristics to meet the actual needs of the spectrum allocation in cognitive radio networks. In addition, by combining the advantages of the methods of AHP and coloring theory, the proposal not only improves the efficiency of spectrum allocation, but also satisfies the requirements of the overall benefits of cognitive radio networks.
2 Spectrum Allocation
2.1 Spectrum Selection Based on AHP In spectrum selection, secondary users always want to switch to the spectrum with high bandwidth, low delay, low jitter and packet loss rate, etc. Therefore, this paper selects four indexes of bandwidth, delay, jitter and packet loss rate as the judgment criterion of spectrum selection. During the course, secondary users should also take their own preferences into account. When the number of considering factors become increasing, secondary users will struggle to make a rational choice by qualitative analysis. We thus introduce the method of AHP to analyze this problem. The selection problem can decomposed into three levels as shown in Fig.
In more complex environment, more evaluation criteria can be introduced to make it closer to realize.
Fig. 1.
Hierarchical graph of optimal spectrum decision based on AHP algorithm
To compare the impacts of factors C , C n of one layer on a factor
1 2 , · · · , C
S i of another layer, such as the importance of different choice criterions on final channel selection, it is essential to make a comparison between only two factors rather than multiple factors at the same time. Selecting two factors C i and C j each time, we use a ij to denote the impact ratio of C i and C j on S i , all the
10 J. Shi et al.
1 A = (a ij ) n×n , a ij > 0, a ij = . (1) a ji We can rewrite
as ⎛ ⎞ ⎜ ⎟
ω
1 /ω 1 ω
1 /ω
2 · · · ω 1 /ω n ⎜ ⎟ω
2 /ω 1 ω
2 /ω
2 · · · ω 2 /ω nA = , (2) ⎜ ⎟ ⎝ . .. ⎠ .. .. ..
. . . ω n /ω ω n /ω n /ω n
1 2 · · · ω n T where ω = (ω , ω is the weighted coefficient vector satisfying ω i = 1.
1 2 , · · · ) i =1
Thus, as mentioned above, proper selection of the weighted coefficients in the secondary user’s utility function is significant. In subsequent simulation tests, we will provide numerical results to testify the effects of different parameter selection in spectrum trading.
2.2 Single Hierarchical Arrangement and Consistency Check Single hierarchical arrangement refers to the same level of the corresponding factors for the relative importance of the upper level of a factor ranking weight, it can be obtained by normalizing the eigenvector (weighted vector) W of the largest eigenvalue λ max of judgment matrix A. So the essence is to calculate the weight vector. The calculation of the weight vector has the characteristic root method, the sum method, the root method, the power method and so on. In this paper, we use the sum method. The steps of the sun method are as follows:
n 1. Normalize each column vector of A : ˜ W ij = a ij / a ij . i =1 n
˜ 2. Make summation for each row of A : ˜ W ij = a i / W i,j .
j =1 n
˜
3. Normalize above matrix vector W = ˜ W i / ˜ W ij = a i / W i , then obtain W =
i =1 T (W , W n ) as weight vector.
1 2 , · · · , W 4. Calculate AW . n i (AW )
1
5. Calculate λ max = , this is the approximate value of the largest
n W i i =1 eigenvalue.
The procedure for the consistency check is as follows: Step 1: Calculate the consistency index (CI).
λ max − n
CI = . (3) n − 1 Step 2: Seek table to determine the corresponding random index (RI).
Spectrum Allocation in Cognitive Radio Networks by Hybrid AHP
11 Table 1. Random index RI Matrix order 1 2 3
4
5
6
7
8
9
10 RI 0 0 0.58 0.9 1.12 1.24 1.32 1.41 1.45 1.40
Step 3: Calculate the consistency ratio (CR) and make judgments.
CR = CI RI
. (4) Fig. 2.
Diagram of greedy spectrum allocation algorithm combined with AHP
12 J. Shi et al.
consistency requirement, the judgment matrix should be revised again. To test the consistency. We take matrix A as an example. First, we use
to calcu-
late the consistency index CI = 0.0083, and then obtain RI = 0.9 via Table Finally, due to the fact that CR = 0.0092 < 0.1, we can conclude that A is verified to pass consistency check.
2.3 Improved Spectrum Allocation Algorithm Model According to the analysis above-mentioned, by selecting the optimal spectrum and then using the graph algorithm to allocate idle spectrum, we can make full use of spectrum resources, enhance the spectrum utilization and improve the overall efficiency.
Combining the distributed greedy algorithm and the method of AHP, the spectrum decision diagram is shown in Fig.
The improved spectrum allocation algorithm can be described as follows: After initializing the system and updating the node information, according to the measured values of bandwidth, delay, jitter and packet loss rate of each spectrum, the hierarchical structure is established by using AHP; Set up the corresponding comparison matrix, and obtain the spectrum efficiency; Finally, the selected optimal spectrum is allocated using the coloring algorithm.
3 Numerical Results
In the simulation process, we suppose that there are six primary users in given region of 1000 m × 1000 m, and the initial number of idle channels is 10. In this simulation environment, the total network benefits and user fairness of the two algorithms and their improved algorithms are analyzed and compared. Using
4 WDGA AHP-DGA DGA
3.5
3
2.5 Network efficiency
2
1.5
40
60 80 100 120 Number of secondary users
Spectrum Allocation in Cognitive Radio Networks by Hybrid AHP
13
4 WDGA AHP-DGA DGA
3.5
3
2.5 Fitness
2
1.5
1
40
60 80 100 120 Number of secondary users Fig. 4.
Changing curves of the three algorithms in greedy mode
U (R) to measure the network efficiency, with the variance to measure the fair- ness between users. In the simulation, randomly generate the network topology diagram. Furthermore, we randomly set the values in available spectrum matrix L , interference matrix C and utility matrix B within [0, 1].
The parameters including bandwidth, delay, jitter and packet loss rate meet the data transmission standard of wireless networks proposed by ITU-T
our proposed method using AHP and distributed greed algorithm
(AHP-DGA) is compared with the results obtained by original distributed greedy algorithm (DGA) and weighted distributed greedy algorithm (WDGA). It can be concluded that AHP-DGA can receive high network efficiency and decent network fairness.
References
1. Fadeel, K.Q.A., Elsayed, D., Khattab, A., Digham, F.: Dynamic spectrum access for primary operators exploiting LTE-A carrier aggregation. In: IEEE ICNC, pp.
143–147 (2015)
2. Li, F., Tan, X., Wang, L.: A new game algorithm for power control in cognitive
radio networks. IEEE Trans. Veh. Technol. 60(9), 4384–4392 (2011)
3. Liu, X., Jia, M., Tan, X.: Threshold optimization of cooperative spectrum sensing
in cognitive radio network. Radio Sci. 48(1), 23–32 (2013)
4. Wang, W., Liu, X.: List-coloring based channel allocation for open-spectrum wireless
networks. In: IEEE VTC-Fall, pp. 690–694 (2005)
5. Liu, Y., Jiang, M., Tan, X., et al.: Maximal independent set based channel allo-
cation algorithm in cognitive radios. In: IEEE Youth Conference on Information, Computing and Telecommunication, pp. 78–81 (2009)14 J. Shi et al.
6. Bao, Y., Wang, S., Yan, B., et al.: Research on maximal weighted independent set-
based graph coloring spectrum allocation algorithm in cognitive radio networks. In: Proceedings of the International Conference on Communications, Signal Processing and Systems, pp. 263–271 (2016)7. ITU-T (2016).
8. Li, M.: Research of cognitive radio spectrum allocation algorithm based on graph
theory. Southwest Jiaotong University, pp. 45–47 (2012)
Spectrum Pricing in Condition of Normally
Distributed User Preference
1 Zhejiang University of Technology, Hangzhou 310023, Zhejiang, China
{liwang2002,fenglzj}@zjut.edu.cn, 478708892@qq.com, 814120631@qq.com