Communications and Networking - Part 1

Molisch University of Southern California, USA Jun Zheng Southeast University Zhi-Feng Zhao Zhejiang University, China Organizing Committee General Chairs Yunjie Liu Academician of Chine

11th EAI International Conference, ChinaCom 2016

Chongqing, China, September 24–26, 2016

Proceedings, Part I

209

for Computer Sciences, Social Informatics

University of Florida, Florida, USA

Xuemin Sherman Shen

University of Waterloo, Waterloo, Canada

Liqiang Zhao (Eds.)

Lecture Notes of the Institute for Computer Sciences, Social Informatics

and Telecommunications Engineering

DOI 10.1007/978-3-319-66625-9

Library of Congress Control Number: 2017953406

© 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, speci fically 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 speci fic 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 af filiations.

Printed on acid-free paper

This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG

The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

On behalf of the Organizing Committee of the 11th EAI International Conference onCommunications and Networking in China (ChinaCom 2016), we would like to wel-come you to the proceedings of this conference ChinaCom aims to bring togetherinternational researchers and practitioners in networking and communications underone roof, building a showcase of these fields in China The conference is beingpositioned as the premier international annual event for the presentation of original andfundamental research advances in thefield of communications and networks.

ChinaCom 2016 was jointly hosted by Chongqing University of Posts andTelecommunications and Xidian University during September 24–26, 2016 Theconference received 181 paper submissions Based on peer reviewing, 107 papers wereaccepted and presented at the conference We thank all the Technical Program Com-mittee (TPC) members and reviewers for their dedicated efforts

ChinaCom 2016 featured six keynote speeches, four invited talks, and a hensive technical program offering numerous sessions in wireless, networks, andsecurity, etc About 150 experts and scholars from more than 10 countries and regionsincluding China, the USA, Canada, Singapore, etc., attend this year’s conference inChongqing

compre-As the youngest municipality of China, Chongqing has become the largest industrialand economic center of the upper Yangtze area Renowned as the Mountain City andfamous for its beautiful and unique spots, Chongqing is a popular destination fortravelers from all over the world

We hope youfind reading the papers in this volume a rewarding experience

Yunjie Liu

Steering Committee

Imrich Chlamtac CREATE-NET (Chair)

Hsiao-Hwa Chen National Cheng Kung University, Taiwan

Ya-Bin Ye Huawei Europe Research Cente

Zheng Zhou Beijing University of Posts and Telecommunications,

China

Bo Li Hong Kong University of Science and Technology,

SAR ChinaAndreas F Molisch University of Southern California, USA

Jun Zheng Southeast University

Zhi-Feng Zhao Zhejiang University, China

Organizing Committee

General Chairs

Yunjie Liu Academician of Chinese Academy of Engineering,

China UnicomYanbin Liu Vice-president, Chongqing University of Posts

and Telecommunications, ChinaTPC Chairs

Weixiao Meng Harbin Institute of Technology, China

Liqiang Zhao Xidian University, China

Qianbin Chen Chongqing University of Posts and Telecommunications,

ChinaLocal Chairs

Zufan Zhang Chongqing University of Posts and Telecommunications,

ChinaJiangtao Luo Chongqing University of Posts and Telecommunications,

ChinaHongxin Tian Xidian University, China

Zhiyuan Ren Xidian University, China

Sponsorship and Exhibits Chair

Qiong Huang Chongqing University of Posts and Telecommunications,

China

Publicity and Social Media Chair

Yang Wang Chongqing University of Posts and Telecommunications,

ChinaWeb Chair

Ting Zhang Chongqing University of Posts and Telecommunications,

ChinaPublication Chair

Rong Chai Chongqing University of Posts and Telecommunications,

ChinaConference Manager

Barbara Fertalova (EAI, European Alliance for Innovation)

TPC Chairs of Chinacom 2016

TPC Chairs

Weixiao Meng Harbin Institute of Technology, China

Qianbin Chen Chongqing University of Posts and Telecommunications,

ChinaLiqiang Zhao Xidian University, China

Symposium Chairs

Future Internet and Networks Symposium

Huaglory Tianfield Glasgow Caledonian University, UK

Guofeng Zhao Chongqing University of Posts and Telecommunications,

ChinaMobile and Wireless Communications Symposium

Lin Dai City University of Hong Kong, SAR China

Yunjian Jia Chongqing University, China

Optical Networks and Systems Symposium

Xingwen Yi University of Electronic Science and Technology of China,

ChinaHuanlin Liu Chongqing University of Posts and Telecommunications,

China

IoT, Smart Cities, and Big Data Symposium

Shensheng Tang Missouri Western State University, USA

Wee Peng Tay Nanyang Technological University, Singapore

Rong Yu Guangdong University of Technology, China

Security Symposium

Qing Yang Montana State University, USA

Yi Qian University of Nebraska Lincoln, USA

Jun Huang Chongqing University of Posts and Telecommunications,

ChinaTechnical Program Committee

Rong Chai Chongqing University of Posts and Telecommunications,

ChinaHongbin Chen Guilin University of Electronic Technology, ChinaZhi Chen University of Electronic Science and Technology of ChinaPeter Chong Nanyang Technological University, Singapore

Dezun Dong National University of Defense Technology, ChinaWei Dong Zhejiang University, China

Jun Fang University of Electronic Science and Technology of ChinaZesong Fei Beijing Institute of Technology, China

Feifei Gao Tsinghua University, China

Ping Guo Chongqing University, China

Guoqiang Hu Nanyang Technological University, Singapore

Tao Huang Beijing University of Posts and Telecommunications,

ChinaXiaoge Huang Chongqing University of Posts and Telecommunications,

ChinaFan Li Beijing Institute of Technology, China

Zhenyu Li Institute of Computing Technology, Chinese Academy

of Sciences, ChinaHongbo Liu Indiana University-Purdue University Indianapolis, USAHongqing Liu Chongqing University of Posts and Telecommunications,

ChinaJiang Liu Beijing University of Posts and Telecommunications,

ChinaQiang Liu University of Electronic Science and Technology of China,

ChinaWenping Liu Hubei University of Economic, China

Rongxing Lu Nanyang Technological University, Singapore

Yilin Mo Nanyang Technological University, Singapore

Jianquan Ouyang Xiangtan University, China

Tian Pan Beijing University of Posts and Telecommunications,

China

Mugen Peng Beijing University of Posts and Telecommunications,

ChinaBin Shen Chongqing University of Posts and Telecommunications,

ChinaYan Shi Beijing University of Posts and Telecommunications,

ChinaGongpu Wang Beijing Jiaotong University, China

Lin Wang Yanshan University, China

Yang Wang Chongqing University of Posts and Telecommunications,

China

Renchao Xie Beijing University of Posts and Telecommunications,

ChinaChangyou Xing PLA University of Science and Technology, ChinaChengwen Xing Beijing Institute of Technology, China

Chuan Xu Chongqing University of Posts and Telecommunications,

ChinaFan Yang Beijing University of Posts and Telecommunications,

ChinaQinghai Yang Xidian University, China

Zhe Yang Northwestern Polytechnical University

Guangxing Zhang Institute of Computing Technology,

Chinese Academy of SciencesJian-Kang Zhang McMaster University, Canada

Jiao Zhang Beijing University of Posts and Telecommunications,

ChinaXiaofei Zhang Nanjing University of Aeronautics and Astronautics, ChinaXing Zhang Beijing University of Posts and Telecommunications,

ChinaYanping Zhang Gonzaga University, USA

Dongmei Zhao McMaster University, Canada

Nan Zhao Dalian University of Technology, China

Yangming Zhao University of Electronic Science and Technology of ChinaSheng Zhou Tsinghua University, China

Zhangbing Zhou China University of Geosciences

Contents – Part I

Technical Sessions

Transceiver Optimization in Full Duplex SWIPT Systems

with Physical Layer Security 3Ruijin Sun, Ying Wang, and Xinshui Wang

Robust Secure Transmission Scheme in MISO Interference Channel

with Simultaneous Wireless Information and Power Transfer 14Chong Xue, Jian Xiao, Sai Zhao, Jingrong Zhou, and Maoxin Tian

An Effective Limited Feedback Scheme for FD-MIMO Based

on Noncoherent Detection and Kronecker Product Codebook 24Lisi Jiang and Juling Zeng

Two-Stage Precoding Based Interference Alignment for Multi-cell

Massive MIMO Communication 34Jianpeng Ma, Shun Zhang, Hongyan Li, and Weidong Shao

MAC Schemes

Adaptive Energy-Saving Mechanism for SMAC Protocol in Wireless

Sensor Network 47Zhou Jieying, Peng Shi, Liu Yinglin, and Huang Shaopeng

A Transmission Rate Optimized Cooperative MAC Protocol

for Wireless Sensor Networks 58Pengfei Zhao, Kai Liu, Feng Liu, and Ruochen Fang

Heterogeneous Control and Data Split Network for Precision

Formation Flying of Distributed Spacecraft 67Haiyan Jiao, Liqiang Zhao, and Xiaoxiao Zhang

A Novel Feedback Method to Enhance the Graphical Slotted ALOHA

in M2M Communications 77

Yu Hanxiao, Jia Dai, Zhang Zhongwei, Sun Ce, Huang Jingxuan,

and Fei Zesong

A Hybrid Automatic Repeat reQuest Scheme Based on Maximum Distance

Separable Codes 87Shangguan Chenglin, Jia Dai, Yang Yanbao, Yu Hanxiao, Sun Ce,

and Fei Zesong

Energy-Efficient Resource Allocation in Distributed Antenna Systems 97Xiaoge Huang, Weipeng Dai, Zhifang Zhang, Qiong Huang,

and Qianbin Chen

Traffic Engineering and Routing Algorithms

Applications of Genetic Algorithms in BGP-Based Interdomain

Traffic Engineering 109Jiyun Yan, Zhenqiang Li, and Xiaohong Huang

MP-SDWN: A Novel Multipath-Supported Software Defined Wireless

Network Architecture 119Chuan Xu, Wenqiang Jin, Yuanbing Han, Guofeng Zhao,

and Huaglory Tianfield

Performance Analysis of Routing Algorithms Based on Intelligent

Optimization Algorithms in Cluster Ad Hoc Network 129Chenguang He, Tingting Liang, Shouming Wei, and Weixiao Meng

Incentive Mechanism for Crowdsensing Platforms Based on Multi-leader

Stackelberg Game 138Xin Dong, Xing Zhang, Zhenglei Yi, and Yiran Peng

Master Controller Election Mechanism Based on Controller Cluster

in Software Defined Optical Networks 148Jie Mi, Xiaosong Yu, Yajie Li, Yongli Zhao, Jie Zhang, Chuan Liu,

and Gang Zhang

Security

Performance Evaluation of Black Hole Attack Under AODV

in Smart Metering Network 159Yanxiao Zhao, Suraj Singh, Guodong Wang, and Yu Luo

An Entropy-Based DDoS Defense Mechanism in Software

Defined Networks 169Yajie Jiang, Xiaoning Zhang, Quan Zhou, and Zijing Cheng

Protecting Location Privacy Through Crowd Collaboration 179Zhonghui Wang, Guangwei Bai, and Hang Shen

A Measurement and Security Analysis of SSL/TLS Deployment

in Mobile Applications 189

Yu Guo, Zigang Cao, Weiyong Yang, and Gang Xiong

A Method for Countering Snooping-Based Side Channel Attacks

in Smart Home Applications 200Jingsha He, Qi Xiao, and Muhammad Salman Pathan

Coding Schemes

FPGA-Based Turbo Decoder Hardware Accelerator in Cloud Radio

Access Network (C-RAN) 211Shaoxian Tang, Zhifeng Zhang, Jun Wu, and Hui Zhu

Iterative Detection and Decoding for Spatially Coupled Multiuser

Data Transmission 221Xiaodan Wang, Sijie Wang, Zhongwei Si, Zhiqiang He, Kai Niu,

and Chao Dong

Two Degree Forest Based LT Codes with Feedback 232Liang Liu and Feng Liu

Joint Spatial Diversity and Network Coding in Satellite Communications 242Cui-Qin Dai, Qingyang Song, Lei Guo, and Nan-Nan Huang

Interference Alignment in Cognitive Relay Networks Under CSI Mismatch 254Weiwei Yang, Tao Zhang, Yueming Cai, and Dan Wu

Joint User Grouping and Antenna Selection Based Massive MIMO

Zero-Forcing Beamforming 264Wang Qian, Hua Quan, Zhou Yingchao, and Shen Bin

Relay Systems

Utility-Based Resource Allocation in OFDMA Relay Systems

with Half-Duplex Transmission 277Huanglong Teng, Binjie Hu, Hongming Yu, Miao Cui,

and Guangchi Zhang

Joint Time Switching and Power Allocation for Secure Multicarrier

Decode-and-Forward Relay Systems with Wireless Information

and Power Transfer 285Xiancai Chen, Gaofei Huang, Yuan Lin, Zijun Liang, and Jianli Huang

Joint Relay Processing and Power Control for Two-Way Relay Networks

Under Individual SINR Constraints 295Dongmei Jiang, Balasubramaniam Natarajan, and Haisheng Yu

Capacity Region of the Dirty Two-Way Relay Channel to Within

Constant Bits 305Zhixiang Deng, Yuan Gao, Wei Li, and Changchun Cai

Quality-of-Service Driven Resource Allocation via Stochastic Optimization

for Wireless Multi-user Relay Networks 316Xiao Yin, Yanbo Ma, Qiang Liu, and Wei Su

System Performance Evaluation and Enhancement

LTE System Performance Evaluation for High-Speed Railway Environment

Under Rician Channel 329Lei Xiong, Ru Feng, and Ting Zhou

A First Look at Cellular Network Latency in China 339Xinheng Wang, Chuan Xu, Wenqiang Jin, and Guofeng Zhao

Rate-Splitting Non-orthogonal Multiple Access: Practical Design

and Performance Optimization 349Xinrui Huang, Kai Niu, Zhongwei Si, Zhiqiang He, and Chao Dong

Improved Proportional Fair Scheduling Mechanism

with Joint Gray-Mapping Modulation for NOMA 360Jing Guo, Xuehong Lin, and Zhisong Bie

Hybrid Interleaved-PTS Scheme for PAPR Reduction in OFDM Systems 370Lingyin Wang

Coverage Probability and Data Rate of D2D Communication Under

Cellular Networks by Sharing Uplink Channel 380Tianyu Zhang, Jian Sun, Xianxian Wang, and Zhongshan Zhang

Optical Systems and Networks

A Novel OFDM Scheme for VLC Systems Under LED

Nonlinear Constraints 393Lingkai Kong, Congcong Cao, Siyuan Zhang, Mengchao Li, Liang Wu,

Zaichen Zhang, and Jian Dang

Design and Implementation of Link Loss Forwarding in 100G Optical

Transmission System 403Zhenzhen Jia, Wen He, Chaoxiang Shi, Jianxin Chang, and Meng Gao

425-Gb/s Duo-Binary System over 20-km SSMF Transmission

with LMS Algorithm 412Mengqi Guo, Ji Zhou, Xizi Tang, and Yaojun Qiao

Self-homodyne Spatial Super-Channel Based Spectrum and Core

Assignment in Spatial Division Multiplexing Optical Networks 423

Ye Zhu, Yongli Zhao, Wei Wang, Xiaosong Yu, Guanjun Gao,

and Jie Zhang

Management of a Hub-Spoken Optical Transmission Network with the

Point to Multi Point (P2MP) Topology 431Wen He, Zhenzhen Jia, Chaoxiang Shi, Jianxin Chang, and Meng Gao

Optimal Power Allocations for Full-Duplex Enhanced Visible Light

Communications 440Liping Liang, Wenchi Cheng, and Hailin Zhang

Signal Detection and Estimation

A Novel Bitwise Factor Graph Belief Propagation Detection Algorithm

for Massive MIMO System 453Lin Li and Weixiao Meng

Development of 4 4 Parallel MIMO Channel Sounder for High-Speed

Scenarios 463Dan Fei, Bei Zhang, Ruisi He, and Lei Xiong

Blind Spectrum Sensing Based on Unilateral Goodness of Fit Testing

for Multi-antenna Cognitive Radio System 472Yinghui Ye and Guangyue Lu

Frequency Detection of Weak Signal in Narrowband Noise Based

on Duffing Oscillator 480Shuo Shi, Qianyao Ren, Dezhi Li, and Xuemai Gu

Basis Expansion Model for Fast Time-Varying Channel Estimation

in High Mobility Scenarios 489Xinlin Lai, Zhonghui Chen, and Yisheng Zhao

Robust Power Allocation Scheme in Cognitive Radio Networks 502Hongzhi Wang, Meng Zhu, and Mingyue Zhou

Author Index 513

Contents – Part II

Energy Harvesting Systems

Energy-Efficient Resource Allocation in Energy Harvesting

Communication Systems: A Heuristic Algorithm 3Yisheng Zhao, Zhonghui Chen, Yiwen Xu, and Hongan Wei

Relay Selection Scheme for Energy Harvesting Cooperative Networks 13Mengqi Yang, Yonghong Kuo, and Jian Chen

Dynamic Power Control for Throughput Maximization in Hybrid Energy

Harvesting Node 23Didi Liu, Jiming Lin, Junyi Wang, Hongbing Qiu, and Yibin Chen

Power Allocation Algorithm for Heterogeneous Cellular Networks

Based on Energy Harvesting 33Xiaoyu Wan, Xiaolong Feng, Zhengqiang Wang, and Zifu Fan

Price-Based Power Allocation in Energy Harvesting Wireless

Cooperative Networks: A Stackelberg Game Approach 44Chongyang Li and Xin Zhao

Resource Allocation Schemes (1)

Coverage and Capacity Optimization Based on Tabu Search

in Ultra-Dense Network 57Xin Su, Xiaofeng Lin, Jie Zeng, and Chiyang Xiao

Dynamic APs Grouping Scheme Base on Energy Efficiency in UUDN 67Shanshan Yu, Xi Li, Hong Ji, and Yiming Liu

Virtual Small Cell Selection Schemes Based on Sum Rate Analysis

in Ultra-Dense Network 78

Qi Zhang, Jie Zeng, Xin Su, Liping Rong, and Xibin Xu

System Level Performance Evaluation for Ultra-Dense Networks 88Qianbin Chen, Ya Zhang, and Lun Tang

Green Distributed Power Control Algorithm for Multi-user Cognitive

Radio Networks 97Yinmeng Wang, Jian Chen, Chao Ren, and Huiya Chang

Optimal Channel Selection and Power Control over D2D Communications

Based Cognitive Radio Networks 107

Ya Gao, Wenchi Cheng, Zhiyuan Ren, and Hailin Zhang

Network Architecture and SDN

Research on Load Balancing for Software Defined Cloud-Fog Network

in Real-Time Mobile Face Recognition 121Chenhua Shi, Zhiyuan Ren, and Xiuli He

Applying TOPSIS Method for Software Defined Networking (SDN)

Controllers Comparison and Selection 132Firas Fawzy Zobary

Robust Congestion Control in NFVs and WSDNs with Propagation Delay

and External Interference 142

Xi Hu and Wei Guo

Latency-Aware Reliable Controller Placements in SDNs 152Yuqi Fan, Yongfeng Xia, Weifa Liang, and Xiaomin Zhang

Signal Detection and Estimation (2)

Multiantenna Based Blind Spectrum Sensing via Nonparametric Test 165Guangyue Lu, Cai Xu, and Yinghui Ye

Blind Spectrum Sensing in Cognitive Radio Using Right Anderson

Darling Test 175Yuxin Li, Yinghui Ye, Guangyue Lu, and Cai Xu

A Computationally Efficient 2-D DOA Estimation Approach for

Non-uniform Co-prime Arrays 183Fenggang Sun, Lei Zhao, Xiaozhi Li, Peng Lan, and Yanbo Zi

Low-Complexity MMSE Signal Detection Based on WSSOR Method

for Massive MIMO Systems 193Hua Quan, Silviu Ciocan, Wang Qian, and Shen Bin

Channel Characteristics and User QoS-Aware Handoff Target Spectrum

Selection in Cognitive Radio Networks 203Hadjor David and Rong Chai

Heterogeneous Networks

A Tractable Traffic-Aware User Association Scheme

in Heterogeneous Networks 217Xiaobing Lin, Kun Yang, and Xing Zhang

An Optimal Joint User Association and Power Allocation Algorithm for

Secrecy Information Transmission in Heterogeneous Integrated Networks 227Mingxue Chen, Yuanpeng Gao, Rong Chai, and Qianbin Chen

Energy-Efficient Femtocells Active/Idle Control and Load Balancing

in Heterogeneous Networks 237Xiaoge Huang, Zhifang Zhang, Weipeng Dai, Qiong Huang,

and Qianbin Chen

Energy Efficiency of Heterogeneous Air-Ground Cellular Networks 248Jie Xin, Liqiang Zhao, and Guogang Zhao

Capacity Analysis in the Cognitive Heterogeneous Cellular Networks

with Stochastic Methods 258Yinglei Teng, Mengting Liu, and Mei Song

A Joint Bandwidth and Power Allocation Scheme for Heterogeneous

Networks 268Yujiao Chen, Hong Chen, and Rong Chai

Internet of Things

A Novel Power-Saving Scheduling Scheme in Large Scale

Smart-Grid Networks 281Chen Chen, Lei Liu, Mingcheng Hu, Qingqi Pei, Li Cong,

and Shengda Wang

Preamble Design for Collision Detection and Channel Estimation

in Machine-Type Communication 292Shilei Zheng, Fanggang Wang, and Xia Chen

A Data Dissemination Strategy in SDN Enabled Vehicular Networks 302Chen Chen, Na Li, Yansong Li, Ronghui Hou, and Zhiyuan Ren

On the Minimum the Sum-of-Squares Indicator of a Balanced

Boolean Function 314

Yu Zhou and Zepeng Zhuo

Distributed Framework for Cognitive Radio Based Smart Grid

and According Communication/Power Management Strategies 322Tigang Jiang

Hardware Design and Implementation

Design of a Cooperative Vehicular Platoon System Based

on Zynq/SoC Architecture 335

Yi Wang, Yi Zhou, Wei Li, Gaochao Wang, Lin Ren, and Ruirui Huang

A Multi-mode Coordinate Rotation Digital Computer (CORDIC) 345Lifan Niu, Xiaoling Jia, Jun Wu, and Zhifeng Zhang

FPGA Design and Implementation of High Secure Channel Coding

Based AES 355Mostafa Ahmed Mohamed Sayed, Liu Rongke, and Zhao Ling

IoT-Architecture-Based All-in-One Monitoring System Design

and Implementation for Data Center 367Jinde Zhou, Wenjun Xu, Fan Yang, and Jiaru Lin

Research on Receiving Visible Light Signal with Mobile Phone 378Qiaozhi Yuan, Zhenshan Zhang, Yaojun Qiao, Ke Liao, and HaiHua Yu

Mobility Management

STGM: A Spatiotemporally Correlated Group Mobility Model

for Flying Ad Hoc Networks 391Xianfeng Li and Tao Zhang

Radial Velocity Based CoMP Handover Algorithm in LTE-A System 401Danni Xi, Mengting Liu, Yinglei Teng, and Mei Song

Optimized Traffic Breakout and Mobility Support for WLAN

and Cellular Converging Network 411Gang Liu

Application of Mobile IP in the Space-Ground Network Based

on GEO Satellites 421Feng Liu, Han Wu, and Xiaoshen Xu

Impact of Doppler Shift on LTE System in High Speed Train Scenario 431

Yu Zhang, Lei Xiong, Xuelian Yang, and Yuanchun Tan

SDN and Clouds

Real-Time Fault-Tolerant Scheduling Algorithm in Virtualized Clouds 443Pengze Guo and Zhi Xue

Resource Allocation with Multiple QoS Constraints in OFDMA-Based

Cloud Radio Access Network 453Shichao Li, Gang Zhu, Siyu Lin, Qian Gao, Shengfeng Xu, Lei Xiong,

and Zhangdui Zhong

Energy-Efficient and Latency-Aware Data Placement for Geo-Distributed

Cloud Data Centers 465Yuqi Fan, Jie Chen, Lusheng Wang, and Zongze Cao

Constrained Space Information Flow 475Alfred Uwitonze, Jiaqing Huang, Yuanqing Ye, and Wenqing Cheng

Hybrid Roadside Devices Placement for Advertisement Disseminations

in Vehicular CPS 486Junshan Cui, Peng Li, Dongdong Yue, Yu Jin, Yu Liu, and Qin Liu

Navigation, Tracking and Localization

A Modified LFF Method for Direct P-Code Acquisition

in Satellite Navigation 499Xinpeng Guo, Hua Sun, Hongbo Zhao, and Wenquan Feng

A Dual-Tone Radio Interferometric Tracking System 509Pan Xiao, Yiyin Wang, Cailian Chen, and Xinping Guan

An Efficient Nonparametric Belief Propagation-Based Cooperative

Localization Scheme for Mobile Ad Hoc Networks 519Chaojie Xu, Hui Yu, and Ming Yang

Mutual Coupling Calibration in Super-Resolution Direction Finding

for Wideband Signals 529Jiaqi Zhen, Danyang Qin, and Bing Zhao

Walking Detection Using the Gyroscope of an Unconstrained Smartphone 539Guodong Qi and Baoqi Huang

FMN

Spectrum Access Based on Energy Harvesting with Optimal

Power Allocation 551Jiaying Wu, Weidang Lu, Hong Peng, and Xin Liu

The CEEFQPSK Scheme for Two-Way Relay Communication Systems

with Physical-Layer Network Coding 560Hongjuan Yang, Jinxiang Song, Bo Li, and Xiyuan Peng

A Brief Review of Several Multi-carrier Transmission Techniques

for 5G and Future Mobile Networks 569Zhen-yu Na, Xiao-tong Li, Xin Liu, Zhi-an Deng, and Xiao-ming Liu

RSSI Based Positioning Fusion Algorithm in Wireless Sensor Network

Using Factor Graph 577Wanlong Zhao, Shuai Han, Weixiao Meng, and Zijun Gong

Crowdsourcing-Based Indoor Propagation Model Localization

Using Wi-Fi 587Yongliang Sun, Jian Wang, Wenfeng Li, Rui Jiang,

and Naitong Zhang

Author Index 597

Technical Sessions

Systems with Physical Layer Security

Ruijin Sun, Ying Wang(B), and Xinshui WangState Key Laboratory of Networking and Switching Technology,

Beijing University of Posts and Telecommunications,

Beijing 100876, People’s Republic of China

wangying@bupt.edu.cn

Abstract To meet the requirements of energy saving, high security and

high speed for the next generation wireless networks, this paper tigates simultaneous wireless information and power transfer (SWIPT)

inves-in full duplex systems takinves-ing the physical layer security inves-into account.Specifically, we consider a full duplex wireless system where a full duplexbase station (FD-BS) communicates with one downlink user and oneuplink user simultaneously, and one idle user also scavenges the radio-frequency (RF) energy broadcasted during the communication for futureuse Since the idle user has great potential to intercept the downlink infor-mation, we assume that FD-BS exploits the artificial noise (AN), which

is another energy source to idle user, to prevent it The imperfect interference cancellation at the FD-BS is considered and the zero forcing(ZF) receiver is adopted to cancel the residual self-interference Then,the optimal transmitter design at FD-BS are derived to maximize theweighted sum rate of downlink secure and uplink transmission, subject

self-to constraints that the transmission power at FD-BS is restricted andthe minimal amount of harvested energy at idle user is guaranteed Theperfect full duplex and half duplex schemes are also introduced for com-parison Extensive simulation results are given to verify the superiority

of our proposed full duplex scheme

Recently, with the exponential surge of energy consumption in wireless nication, green communications have received much attention from both indus-try and academic As a promising technology towards green communications,harvesting the ambient radio-frequency (RF) energy can prolong the lifetime ofenergy-constrained wireless networks More importantly, scavenging energy fromthe far-field RF signal transmission enables simultaneous wireless informationand power transfer (SWIPT) [1] Typically, there exist fundamental tradeoffsbetween harvested energy and received information rate Many works focused

commu-c

 ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2018

Q Chen et al (Eds.): ChinaCom 2016, Part I, LNICST 209, pp 3–13, 2018.

on downlink SWIPT systems where a transmitter serves two kinds of receivers,i.e., information decoding receivers (IRs) and energy harvesting receivers (ERs).Based on this scenario, joint information beamforming for IRs and energy beam-forming for ERs were investigated [2,3] In particular, to meet the different powersensitivity requirements of energy harvesting (EH) and information decoding(ID) (e.g., −10 dBm for EH versus −60 dBm for ID), a location-based receiver

scheduling scheme was proposed in [3], where ERs need to be closer to the mitter than IRs This scheme indeed facilitates the energy harvesting at ERssince they always have better channels due to distance-dependent attenuation.However, this receiver scheduling scheme may also increase the susceptibility

trans-to eavesdropping, because that ERs, the potential eavesdroppers, can more ily overhear the information sent to IRs In traditional communication networkswithout energy harvesting, this security issue can be addressed from the physicallayer perspective, by transmitting additional artificial noise (AN) to degrade thechannel of eavesdroppers [4] When it comes to downlink SWIPT systems, thepower stream for energy supply can naturally serve as AN to prevent eavesdrop-ping Thus, secure communication in downlink systems with SWIPT was studied[5] In [5], Liu et al presented a system secrecy rate maximization problem and

eas-a weighted sum-heas-arvested-energy meas-aximizeas-ation problem vieas-a the joint design ofinformation and energy beamforming

Apart from energy saving and high information security, high informationspeed is also a main objective of next generation wireless communications To thisend, full duplex, which has the potential to double the system spectral efficiency,has aroused researchers’ wide concern The benefits are intuitively brought byallowing signal transmission and reception at the same time and the same fre-quency Recently, the strong self-interference (SI) that full duplex systems sufferfrom can been greatly suppressed via the effective self-interference cancellation(SIC) techniques, such as antenna separation, analog domain suppression anddigital domain suppression [6] Consequently, a majority of researches on fullduplex systems have been investigated, including the re-designed SIC [7] andspectral efficiency analysis [8]

In order to meet the requirements of energy saving, high security as well ashigh speed for the next generation wireless networks, in this paper, we studyfull duplex SWIPT systems with the physical layer security Specifically, weconsider a full duplex wireless system where the full duplex base station (FD-BS) communicates with one downlink user and one uplink user simultaneously,and one idle user scavenges the RF energy broadcasted during the communica-tion for future use Since the idle user has the great potential to intercept thedownlink information, we assume that FD-BS exploits the artificial noise (AN),which is another energy source to idle user, to prevent it Similar to full duplexcommunication systems [8], the proposed secrecy SWIPT full duplex scenario

is also subject to the practical issue of imperfect SIC at the FD-BS To reducethe computational complexity, the optimal transmitter design with the fixedzero forcing (ZF) receiver at FD-BS are derived to maximize the weighted sumrate of downlink secure and uplink transmission, subject to constraints that the

transmission power at FD-BS is restricted and the minimal amount of harvestedenergy at idle user is guaranteed The objective function of original non-convexoptimization is transformed into a linear fractional form by introducing an non-negative parameter Then, by applying Charnes-Cooper transformation, semi-definite programming (SDP) and the bi-search method, the optimal parameter

as well as optimal transmitter design is achieved Simulation results are given toverify the superiority of our proposed full duplex scheme

The remainder of the paper is organized as follows In Sect.2, system modeland problem formulation are introduced In Sect.3we state the ZF receiver basedoptimal transmitter optimization Finally, the simulation results are presented

in Sect.4before Sect.5 concludes the paper

Notation: Bold lower and upper case letters are used to denote column vectors

and matrices, respectively The superscripts HT, HH, H−1 are standard

trans-pose, (Hermitian) conjugate transpose and inverse of H, respectively rank(S) and Tr(S) denote the rank and trace of matrix S, respectively S 0( 0) means

that matrix S is positive semidefinite (positive definite).

Considering a full-duplex system where one FD-BS, one uplink user (U U), one

downlink user (U D ) and one idle user (U I) are included, as illustrated in Fig.1

The FD-BS concurrently communicates with U D in the downlink and U U in theuplink Meanwhile, the idle user scavenges the RF energy broadcasted during the

communication Assume that FD-BS has N = N T + N R antennas, of which N T are used for downlink transmission and N Rare used for uplink receiving Otherusers in the system all have a single antenna due to the hardware limitation.Suppose that FD-BS knows all the channel state information (CSI) The idleuser also feedbacks its CSI to FD-BS for the purpose of harvesting more energy

Fig 1 System Model

In order to facilitate energy harvesting, the idle user is assumed to bedeployed in more proximity to the FD-BS than the downlink and uplink user

Thus, signals transmitted by FD-BS is a dominant part of the signals received

at idle user It becomes more easier for the vicious idle user to eavesdrop theinformation sent by FD-BS Consequently, in this paper, we mainly prevent theeavesdropping in the downlink channel

To prevent the eavesdropping, AN is adopted at FD-BS The transmit sage broadcasted by FD-BS is then given as

where sD ∈ C N T ×1 is the useful signal vector for U

D and sD ∼ CN (0, S) with

the covariance matrix S 0 v ∈ C N T ×1 is the AN vector with v∼ CN (0, V)

and V 0 Note that AN also provides another energy source for the idle user.

The data symbol sent by U U is s U ∼ CN (0, 1) and its transmission power is

P U Hence, denote the message sent by U U as

I ∈ C N T ×1 denote the channel vector from FD-BS to

U D and U I , respectively g I represents the complex channel coefficient from U U

to U I z D ∼ CN (0, σ2

Z ) and z I ∼ CN (0, σ2

Z) are the corresponding background

noise at U D and U I, respectively In this paper, we assume that the scheduled

U D and U U are far from each other and thus ignore the co-channel interference

(CCI) from U U to U D Note that the third term in (4) also plays as noise toavoid malicious eavesdropping and thus we call it as uplink noise (UN)

The received signal to interference plus noise ratio (SINR) at U D is given by

Thus, the achievable secrecy rate at downlink user U Dis represented as

RsecD (S, V) = log2(1 + γ D)− log2(1 + γ I)

Meanwhile, the amount of harvested energy at U I is expressed as

E = ζ

hH

where 0 < ζ  1 is the RF energy conversion efficiency

Next, for the uplink channel, we denote the received signal vector at FD-BSas

ZIN R) is the noise vector wR ∈ C N R ×1is the receive beamforming

at FD-BS and the matrix HSI ∈ C N R ×N T is the self-interference (SI) channelfrom the transmit antennas to the receive antennas at FD-BS

Thus, the uplink channel information rate and SINR can be respectivelygiven by

R U (S, V, w R) = log2(1 + γ U) (10)and

From (3), (4) and (9), we observe that the secure downlink and uplink

trans-mission are coupled by the SI and the UN Since we assume that U I is interested

in the information of FD-BS, the downlink secrecy rate and the uplink rate aretwo main objectives we desire to optimize In order to achieve a tradeoff betweenthem, the weighted sum rate of the secure downlink and uplink transmission,which is a very common and useful method to address the multi-objective opti-mization problem, are maximized in this paper In particular, the problem isexpressed as

where P BS is the allowable transmission power at the FD-BS and e2 is the

minimal amount of energy harvested by idle user U I w D and w U are the positivedownlink and uplink weighted factors, respectively (12b) and (12c) are the power

constraint at FD-BS and harvested energy constraint at U I, respectively.According to [2], the feasible condition of problem P1 is e2 

ζ

P BS h I 22+ P U |g I |2 Throughout this paper, we consider the non-trivial casewhere the positive downlink secrecy rate is achievable

3 Optimal Transmitter Design with ZF Receiver

Under the assumption of N R > N T, ZF receiver is designed to cancel the SI

at the FD-BS perfectly, i.e., wH RHSISHH

where U∈ C N T ×N T and V ∈ C N R ×N R are unitary matrices, Λ is a N T × N R

rectangular diagonal matrix In addition, ˆV ∈ C N R ×N T and ˜V ∈ C N R ×(N R −N T)

is made up of the first N T and the last N R − N T right singular vectors of HH SI,

respectively Note that ˜ V with ˜ VHV = I forms an orthogonal basis for the null ˜ space of HH SI Hence, to satisfy HH SIwR= 0, wR is expressed as:

Notice that design of ˜ wR is only related to uplink transmission after SI

cancellation It can be shown that to maximize the uplink rate, ˜ wR should

be aligned to the same direction as the equivalent channel ˜ VHgU, i.e., ˜ w∗

Next, we only focus on the maximization of secrecy downlink rate According

to [14], the original problem P1 without uplink rate is equivalent to following P2 with the optimal SINR constraint γ i

S,V

Tr(HDS) Tr(HD V) + σ2

I , and γ i is an introducing positive variable.

However, it is still a non-convex optimization problem due to the linear fractionalobjective function

From Charnes-Cooper transformation [10], we define ¯S = ρS, ¯ V = ρV and

rewrite the problemP2 in terms of ¯S and ¯V.

This is a convex SDP problem1 and hence can be solved by CVX [11]

By denoting its objective value as h(γ i), the original problem P1 without

uplink rate becomes max

Proof: We first prove that h(γ i ) is concave in γ i Let λ, μ, ν, θ denote the dual

variables of the corresponding constraints in problem P2.1, respectively Then

the Lagrangian function of problemP2.1 is given by

L(¯ S, ¯ V, ρ, λ, μ, ν, θ, γ i) = Tr(A¯ S) + Tr(B ¯V) + ηρ + λ (18)where

duality holds Thus, h(γ i) = min

λ,μ,ν,θ g(λ, μ, ν, θ, γ i ) It is easily verified that h(γ i)

is a point-wise minimum of a family of affine function and hence concave for

γ i > 0 [12]

1 It has been proved in [5] that, there exist ¯ S∗ and ¯ V∗ which satisfy rank(¯ S∗) = 1

and rank( ¯ V∗) = 1

Then, we use the definition of quasi-concave to prove that f (γ i) is

quasi-concave The superlevel set of function f (γ i) is{γ i |1 + h(γ i) α(1 + γ i)} which

is a convex set due to the concavity of h(γ i ) So f (γ i) is a quasi-concave function

in γ i and its maximum can be found through a one-dimensional search This

In order to find the optimal γ i , we take the gradient of f (γ i), i.e.,

S∗ , ¯V∗ , ρ ∗ are the optimal primary variables and λ ∗ , μ ∗ , ν ∗ , θ ∗ are the optimal

dual variables for a given γ i , respectively With (18), the gradient of h(γ i) can

Above all, problemP2 can be solved in two steps: (i) Given any γ i > 0, we

first solve the Problem P2.1 to obtain h(γ i ) and f (γ i); (ii) Then, we use the

bisection method to find optimal γ i by using the gradient of f (γ i) Repeat thesetwo procedures until problem converges Detailed steps of proposed algorithmare outlined in Algorithm1 Different from Algorithm1, the global optimizationsolution to the problemP2 can be achieved by Algorithm1

Algorithm 1 SDP based bisection method for problem P2

In this section, computer simulation results are presented Throughout the

sim-ulations, the transmission power of FD-BS is set as P BS = 10 W The number

of antennas at FD-BS is N = 6 We set N T = 2, N R = 4 The uplink usertransmission power is set as 1 W The energy harvesting efficiency is set as 50%and the weighted factors of downlink and uplink transmission are equal to 1 forsimplicity We assume that the noise power is the same and equals to −80 dB.

The channel attenuation from FD-BS to downlink user and uplink user is both

70 dB, and the channel attenuation from FD-BS to idle user is 50 dB Thesechannel entries are independently generated from i.i.d Rayleigh fading with the

respective average power values Moreover, the elements of HSI is assumed to

beCN (0, σ2

SI ), where σ2SI is decided by the capability of the SIC techniques and

is also equivalent to the negative value of self-interference channel attenuation

In addition to the proposed full duplex scheme, the perfect full duplex schemeand the half duplex scheme are also introduced for comparison The perfect fullduplex scheme means that the SI is perfectly canceled by SIC techniques In the

half duplex scheme, all N = 6 antennas are used for data transmission/reception

in 1/2 time slot All results in this section are obtained by averaging over 100independent channel realizations Note that whenever a channel realization or aparameter setting makes the problem infeasible, the achievable sum rate is set

to zero

Iteration index

8 8.5 9 9.5 10 10.5 11

Proposed full duplex

SI =−90 dB, e2=−20 dBm

At first, we illustrate the convergence of proposed Algorithm1in Fig.2with

σ2SI =−90 dB, e2 =−20 dBm Each point on the curves of Fig.2 records theoptimal sum rate achieved at each iteration It is shown that the Algorithm1

converges to the optimal value within several iterations

The impact of the minimum energy requirement on the weighted sum ratefor different schemes are shown in Fig.3 with σ SI2 = −90 dB Note that the

achieved weighted sum rate of FD-BS decreases with the increasing of the energydemand What is more, both the proposed full duplex scheme and the perfect

full duplex scheme greatly outperform half duplex scheme when e2< −13 dBm.

However, their performances are slightly worse than that of half duplex when

e2 > −12 dBm The reason is that it is easier for the half duplex scheme with

N = 6 downlink antennas in 1/2 time slot than the full duplex schemes with

-30 -25 -20 -15 -10 -5 0

Minimum energy requirement, e 2 (dBm)

0 2 4 6 8 10 12

Proposed full duplex Perfect full duplex Half duplex

Fig 3 The impact of the minimum energy requirement on the weighted sum rate for

to intercept the downlink information, we have assumed that FD-BS exploitsthe artificial noise (AN), which is another energy source to idle user, to prevent

it The weighted sum rate of the secure downlink and uplink transmission hasbeen maximized given the maximal allowable transmission power and the mini-mal harvested energy requirement Extensive numerical experiments have beencarried out to evaluate the sum rate performance of our proposed schemes

Acknowledgment This work was supported by National Natural Science

Founda-tion of China (Project 61431003, 61421061) and NaFounda-tional 863 Project 2014AA01A705

References

1 Varshney, L.R.: Transporting information and energy simultaneously In: IEEEInformation Theory (ISIT), pp 1612–1616 (2008)

2 Zhang, R., Ho, C.K.: MIMO broadcasting for simultaneous wireless information

and power transfer IEEE Trans Wirel Commun 12(5), 1989–2001 (2013)

3 Xu, J., Liu, L., Zhang, R.: Multiuser MISO beamforming for simultaneous wireless

information and power transfer IEEE Trans Signal Process 62(18), 4798–4810

(2014)

4 Goel, S., Negi, R.: Guaranteeing secrecy using artificial noise IEEE Trans Wirel

Commun 7(6), 2180–2189 (2008)

5 Liu, L., Zhang, R., Chua, K.-C.: Secrecy wireless information and power transfer

with MISO beamforming IEEE Trans Signal Process 62(7), 1850–1863 (2014)

6 Duarte, M., Sabharwal, A.: Full-duplex wireless communications using off-the-shelf:feasibility and first results In: IEEE Signals, Systems and Computers (ASILO-MAR), pp 1558–1562 (2010)

7 Ahmed, E., Eltawil, A.M., Sabharwal, A.: Self-interference cancellation with phasenoise induced ICI suppression for full-duplex systems In: IEEE Global Communi-cations Conference (GLOBECOM), pp 3384–3388 (2013)

8 Wang, Y., Sun, R., Wang, X.: Transceiver design to maximize the weighted sum

secrecy rate in full-duplex SWIPT systems IEEE Signal Process Lett 23(6), 883–

887 (2016)

9 Zhang, L., Zhang, R., Liang, Y.-C., Xin, Y., Cui, S.: On the relationship betweenthe multi-antenna secrecy communications and cognitive radio communications

IEEE Trans Commun 58(6), 1877–1886 (2010)

10 Charnes, A., Cooper, W.W.: Programming with linear fractional functionals Nav

Res Logist Q 9(3), 181–186 (1962) quarterly

11 Grant, M., Boyd, S.: CVX: Matlab software for disciplined convex programming,

12 Boyd, S., Vandenberghe, L.: Convex Optimization Cambridge University Press,Cambridge (2004)

Interference Channel with Simultaneous Wireless Information and Power Transfer

Chong Xue1(B), Jian Xiao1, Sai Zhao2, Jingrong Zhou3, and Maoxin Tian1

Guangzhou 510006, Guangdong, Chinaxuech6@mail2.sysu.edu.cn

Guangzhou 510006, Guangdong, China

Abstract Considering simultaneous wireless information and power

transfer (SWIPT), we investigate robust secure transmission scheme intwo-user multiple-input-single-output interference channels, where chan-nel uncertainties are modeled by worst-case model Our objective is tomaximize the worst-case sum secrecy rate under individual transmitpower constraints and worst-case energy harvest (EH) constraints Wepropose an alternative optimization (AO) based algorithm to solve therobust secure transmission problem, and we can obtain a closed formsolution in the process of AO algorithm Simulation results demonstratethat our proposed robust secure transmission scheme has significant per-formance gain over the non-robust one

Keywords: Simultaneous wireless information and power transfer

Multiple-input-single-output (MISO)

Recently, a unified study on simultaneous wireless information and power fer (SWIPT) has drawn significant attention, which is not only theoreticallyintricate but also practically valuable for enabling both the wireless data andwireless energy access to mobile terminals at the same time For two-user single-input-single-output (SISO), MISO, and multiple-input-multiple-output (MIMO)interference channels (IFCs), SWIPT schemes were invested in [1,2]

trans-Due to the openness of wireless transmission medium and the inherent domness of wireless channel, radio transmission is vulnerable to attacks fromunexpected eavesdroppers [3,4] Secure communications in MISO SWIPT sys-tems were derived in [4,5] where perfect channel state information (CSI) wasconsidered In practice, it is difficult to obtain perfect CSI because of channel

ran-c

 ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2018

Q Chen et al (Eds.): ChinaCom 2016, Part I, LNICST 209, pp 14–23, 2018.

estimation and quantization errors Considering the worst-case channel tainties, the robust secure beamforming scheme with SWIPT in MISO channelswas proposed in [6].

uncer-In this paper, we investigate robust secure transmission scheme in two-usermultiple-input-single-output (MISO) interference channels, where channel uncer-tainties are modeled by worst-case model Our objective is to maximize theworst-case sum secrecy rate under individual transmit power constraints andworst-case energy harvest (EH) constraints The formulated optimization prob-lem is nonconvex and we propose an alternative optimization (AO) based algo-rithm to solve the robust secure transmission problem, and we can obtain aclosed form solution in the process of AO algorithm

A System Model

Consider a two-user MISO IFC system with SWIPT which consists of two mitters, two ID receivers, a eavesdropper and K EH receivers Each transmitter

trans-is equipped with N antennas Each energy receiver and the eavesdropper are

equipped with single antenna Each ID receiver decodes the information sentfrom its correspondence transmitter whereas each EH receiver harvests energyfrom both transmitters The eavesdropper decodes the information sent fromthe two transmitters Denote the channel responses from transmitter i to ID

receiverj, energy receiver k and eavesdropper e as h ij ∈ C N × 1, g ik ∈ C N × 1

and hie ∈ C N × 1.

Denote the confidential signal sent by transmitteri as x i ∈ C N×1,i ∈ {1, 2},

whereE[xix† i] = I and E[xix† j] = I fori = j Thus, the received signals at the ID

receiverj and the eavesdropper e, denoted as y j andy e, respectively, are

where n j ∼ CN (0, σ2I) and n e ∼ CN (0, σ2I) are the additive Gaussian noises

at the ID receiverj and the eavesdropper e, respectively Without loss of

gener-ality, we assume that the noise variance is σ2= 1 in this paper Therefore, the

achievable rate of the ID receiver 1, 2 can be expressed as

I e(x1, x2) = log2(1 + h† 1ex1x†1h1e+ h† 2ex2x†2h2e) (4)

According to [7], the worst-case sum secrecy rate of the system can beexpressed as

I S =I1(x1, x2) +I2(x1, x2 − I e(x1, x2 (5)The transmit power constraint at the transmitteri is

The harvested energy at energy receiverk should be constrained as

ρ(g † 1kx1x†1g1k+ g† 2kx2x†2g2k)≥ Q k (7)whereρ is the EH efficiency that accounts for the loss in energy transducer and

Q k is the threshold of the harvested energy at EH receiver k Without loss of

generality, the EH efficiency is assumed to beρ = 1 in this paper.

B Problem Formulation

We assume that the two transmitters know the imperfect CSI on hij, gik and

hie This assumption is valid because of channel estimation and quantizationerrors In this paper, we model the channel uncertainties by worst-case model as

in [6] which can be expressed as

H ij ={h ij |h ij= ˆhij+Δh ij , Δh † ijVij Δh ij ≤ 1}, (8)

G ik={g ik |g ik= ˆgik+Δg ik , Δg ik † Vik Δg ik ≤ 1}, (9)

H ie={h ie |h ie= ˆhie+Δh ie , Δh † ieVie Δh ie ≤ 1} (10)where ˆhij, ˆgikand ˆhiedenote the estimates of channels hij, gikand hie, respec-tively;Δh ij,Δg ikandΔh ie denote the channel uncertainties; Vij 0, V ik 0

and Vie 0 determine the qualities of CSI.

Considering worst-case channel uncertainties, our objective is to maximizeworst-case sum secrecy rate subject to individual transmit power constraints

at two transmitters and worst-case EH constraints at EH receivers Thus, theoptimization problem is formulated as

whereK = {1, 2, , K} The robust problem (11) is non-convex which is difficult

to solve Thus, we propose an alternative iteration (AO) algorithm to solve theworst-case sum secrecy rate maximization problem

3 Robust Secure Transmission Scheme

It is observed that the optimization problem (11) is a fractional quadraticallyconstrained quadratic (QCQP) problem, which is non-convex and difficult tosolve Employing the semidefinite relaxation method [8], the problem (11) isequivalently rewritten as

In (12a), since η1, η3 are concave and η2, η4, η5 are convex, (12a) is

non-convex In order to deal with (12a), we have the following proposition

Proposition 1 Let a ∈ R 1×1 be a positive scalar and f(a) = −ab

ln 2+log2a+ 1

ln 2.

We have

− log2b = max

and the optimal solution to the right-hand side of (19) is a =1

b Proof Since f(a) is concave, the partial derivative of f(a) with respect to a is

Using Proposition 1, we transformη2, η4, η5 into convex optimization problems

ln 2 (26)

In the following, we propose to decouple (12) into four optimization problemsand employ AO algorithm to iteratively optimize a1, a2, a3, X1 and X2 Our

design concept is based on the fact that for fixed X1and X2the optimal solution

ofa1, a2, a3 can be derived, and vice versa

Given X(n−1)1 and X(n−1)2 which are optimal in the (n − 1)th iteration, we

κ1=tr(ˆh21hˆ†21X(n−1)2 )

tr(X(n−1)2 V−1

21)

(33)

According to the Proposition1and (30)–(33), the closed form of the problem(27) is

tractable, we convert the constraints (39c)–(39i) into linear matrix inequalities(LMIs) [9] equivalently, usingS-Procedure [10].

ApplyingS-Procedure, and introducing slack variables μ1, μ2, μ3, μ4, μ5, μ6

and μ7, the constraints (39c)–(39i) can be equivalently transformed into thefollowing LMIs

ˆ

ϕ = ˆg † 1kX1ˆg1k+ ˆg2k † X2gˆ2k − μ7− μ8− Q k (49)Combing (41)–(46) and (48) and omitting the rank-one constraint, the opti-mization problem (39) can be recast as

max

Wl j ≥ 0, l ∈ {1, , 8}/7, j ∈ {1, , 8} (50c)Obviously, (50) is a convex SDP problem which can be solved by existing soft-ware, e.g., CVX It is noted that (50) is a rank-one relaxation of the originalproblem (39) If the optimal solution of (50) is rank-one, it is also the optimalsolution of the original problem (39) If the rank of the optimal solution of (50) isgreater than 1, we employ the Gaussian randomization (GR) method to generatethe suboptimal rank-one solution