Download the Full Version of textbook for Fast Typing at textbookfull.

com

Data-Driven Wireless Networks: A Compressive

Spectrum Approach Yue Gao

https://textbookfull.com/product/data-driven-wireless-
networks-a-compressive-spectrum-approach-yue-gao/

OR CLICK BUTTON

DOWNLOAD NOW

Download More textbook Instantly Today - Get Yours Now at textbookfull.com

 Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.

Wireless Public Safety Networks 2. A Systematic Approach

1st Edition Daniel Camara

https://textbookfull.com/product/wireless-public-safety-
networks-2-a-systematic-approach-1st-edition-daniel-camara/

textboxfull.com

Monitoring Multimode Continuous Processes: A Data-Driven

Approach Marcos Quiñones-Grueiro

https://textbookfull.com/product/monitoring-multimode-continuous-
processes-a-data-driven-approach-marcos-quinones-grueiro/

textboxfull.com

Design of Video Quality Metrics with Multi Way Data

Analysis A data driven approach 1st Edition Christian
Keimel (Auth.)
https://textbookfull.com/product/design-of-video-quality-metrics-with-
multi-way-data-analysis-a-data-driven-approach-1st-edition-christian-
keimel-auth/
textboxfull.com

Data Visualisation A Handbook for Data Driven Design Andy

Kirk

https://textbookfull.com/product/data-visualisation-a-handbook-for-
data-driven-design-andy-kirk/

textboxfull.com
Data Visualisation: A Handbook For Data Driven Design Andy
Kirk

https://textbookfull.com/product/data-visualisation-a-handbook-for-
data-driven-design-andy-kirk-2/

textboxfull.com

Spectrum Access and Management for Cognitive Radio

Networks 1st Edition Mohammad A Matin (Eds.)

https://textbookfull.com/product/spectrum-access-and-management-for-
cognitive-radio-networks-1st-edition-mohammad-a-matin-eds/

textboxfull.com

Wireless MEMS networks and applications 1st Edition

Uttamchandani

https://textbookfull.com/product/wireless-mems-networks-and-
applications-1st-edition-uttamchandani/

textboxfull.com

Security in Fixed and Wireless Networks Guenter Schaefer

https://textbookfull.com/product/security-in-fixed-and-wireless-
networks-guenter-schaefer/

textboxfull.com

Opportunistic Spectrum Utilization in Vehicular

Communication Networks 1st Edition Nan Cheng

https://textbookfull.com/product/opportunistic-spectrum-utilization-
in-vehicular-communication-networks-1st-edition-nan-cheng/

textboxfull.com
SPRINGER BRIEFS IN
ELEC TRIC AL AND COMPUTER ENGINEERING

Yue Gao
Zhijin Qin

Data-Driven
Wireless
Networks
A Compressive
Spectrum
Approach
123
SpringerBriefs in Electrical and Computer
Engineering

Series editors
Woon-Seng Gan, School of Electrical and Electronic Engineering, Nanyang
Technological University, Singapore, Singapore
C.-C. Jay Kuo, University of Southern California, Los Angeles, CA, USA
Thomas Fang Zheng, Research Institute of Information Technology, Tsinghua
University, Beijing, China
Mauro Barni, Department of Information Engineering and Mathematics, University
of Siena, Siena, Italy
SpringerBriefs present concise summaries of cutting-edge research and practical
applications across a wide spectrum of fields. Featuring compact volumes of 50
to 125 pages, the series covers a range of content from professional to academic.
Typical topics might include: timely report of state-of-the art analytical techniques,
a bridge between new research results, as published in journal articles, and a
contextual literature review, a snapshot of a hot or emerging topic, an in-depth case
study or clinical example and a presentation of core concepts that students must
understand in order to make independent contributions.

More information about this series at http://www.springer.com/series/10059

Yue Gao • Zhijin Qin

Data-Driven Wireless
Networks
A Compressive Spectrum Approach

123
Yue Gao Zhijin Qin
School of Electronic Engineering School of Electronic Engineering
and Computer Science and Computer Science
Queen Mary University of London Queen Mary University of London
London, UK London, UK

ISSN 2191-8112 ISSN 2191-8120 (electronic)

SpringerBriefs in Electrical and Computer Engineering
ISBN 978-3-030-00289-3 ISBN 978-3-030-00290-9 (eBook)
https://doi.org/10.1007/978-3-030-00290-9

Library of Congress Control Number: 2018956141

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2019

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.

This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
How can wireless data influence the
compressive spectrum sensing in windband
cognitive radio and Internet of Things
networks?
Foreword

It is my great pleasure to write a foreword to this fantastic book, Data-Driven

Wireless Networks: A Compressive Sensing Approach, authored by two of my good
friends and long-term collaborators. I first knew the real meaning of “wideband
compressive sensing” several years ago after I discussed with them on research
in cognitive radio. Later on, I read some of their papers in the related topics and
gradually realized that it is a very effective tool in signal processing and wireless
communications. I once thought that it would be great to have a comprehensive
book in the area.
This book addresses spare representation in wireless communications, with
emphasis on the most recently developed compressive sensing-enabled approaches.
It starts from a comprehensive overview of the fundamental principles. Subse-
quently, it introduces the data-driven compressive spectrum sensing in cognitive
radio networks and discusses robust and security issues. The book also contains test
results of various algorithms based on the real-world signals and data collected by
experiments carried out during the TV white space pilot trials.
I would like to emphasize that the authors are very active in the related areas
in the past several years. Their publications have been very well-cited and received
best paper awards. I believe that this book is as good as their other publications.
This book contains most of the important results in the related topics, including
their own. The book is written concisely and clearly. It is an excellent reference for
graduate students, faculty, and practice engineers in the area.

Professor and Director Geoffrey Ye Li

Information Transmission and Processing Lab
School of Electrical and Computer Engineering
Georgia Institute of Technology
Atlanta, GA, USA
June 2018

vii
Preface

In this book, we will discuss the applications of spare representation in wireless

communications, with particular focus on the most recently developed compres-
sive sensing-enabled approaches. With the help of sparsity property, sub-Nyquist
sampling can be achieved in wideband cognitive radio networks by adopting com-
pressive sensing. This book starts from a comprehensive overview of CS principles.
Subsequently, we will present a complete framework for data-driven compressive
spectrum sensing in cognitive radio networks, which is able to provide guarantee on
robustness, low complexity, and security. Particularly, robust compressive spectrum
sensing, low-complexity compressive spectrum sensing, and secure compressive
sensing-based malicious user detection are proposed to address the various issues
in wideband cognitive radio networks. Correspondingly, the real-world signals and
data collected by experiments carried out during TV white space pilot trial enable
data-driven compressive spectrum sensing. The collected data are used to verify our
designs and provide significant insights on the potential of applying compressive
sensing to wideband spectrum sensing. We believe this book will provide readers a
clear picture on how to exploit the compressive sensing to process wireless signals
in wideband cognitive radio networks.

London, UK Yue Gao

June 2018

ix
Acknowledgment

I would like to express my sincere gratitude to my current and former PhD students,
my collaborators, and all the colleagues who contributed to the work and projects
that lead to this book. I would like to thank the UK Engineering and Physical Sci-
ences Research Council (EPSRC) for funding my research projects, e.g., the EPSRC
Fellowship project (EP/R00711X/1).
I would also like to particularly thank our editor as well as all the editorial staff
from Springer in producing this book.
Finally, I would like to thank my family, including my wife, my daughter, and
my son. I cannot make it without their great support.

xi
Contents

Part I Background
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1 Motivations and Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.1.1 Data-Driven Compressive Spectrum Sensing . . . . . . . . . . . . . . . . . . 5
1.1.2 Robust Compressive Spectrum Sensing . . . . . . . . . . . . . . . . . . . . . . . . 5
1.1.3 Secure Compressive Spectrum Sensing . . . . . . . . . . . . . . . . . . . . . . . . 6
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2 Sparse Representation in Wireless Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1 Principles of Standard Compressive Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1.1 Sparse Representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.1.2 Projection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.1.3 Signal Reconstruction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2 Reweighted Compressive Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3 Distributed Compressive Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.4 Compressive Spectrum Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.4.1 Spectrum Sensing Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.4.2 Spectrum Sensing Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.4.3 Compressive Wideband Spectrum Sensing . . . . . . . . . . . . . . . . . . . . . 17
2.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Part II Compressive Spectrum Sensing Algorithms

3 Data-Driven Compressive Spectrum Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.1.1 Related Work. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.1.2 Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.2 Data-Driven Compressive Spectrum Sensing Framework . . . . . . . . . . . . . 25
3.2.1 Iteratively Reweighted Least Square-Based
Compressive Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

xiii
xiv Contents

3.2.2
Non-iteratively Reweighted Least Square-Based
Compressive Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.2.3 Proposed Wilkinson’s Method-Based DTT Location
Probability Calculation Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.3 Numerical Analyses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.3.1 Numerical Analyses on Simulated Signals and Data . . . . . . . . . . 33
3.3.2 Numerical Analyses on Real-World Signals and Data . . . . . . . . . 38
3.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4 Robust Compressive Spectrum Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.1.1 Related Work. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.1.2 Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.2 Robust Compressive Spectrum Sensing at Single User . . . . . . . . . . . . . . . . 45
4.2.1 System Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.2.2 Computational Complexity and Spectrum Usage Analyses . . . 47
4.3 Numerical Analyses for Single User Case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.3.1 Analyses on Simulated Signals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.3.2 Analyses on Real-World Signals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.4 Matrix Completion-Based Robust Spectrum Sensing at
Cooperative Multiple Users . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.4.1 System Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
4.4.2 Denoised Cooperative Spectrum Sensing Algorithm . . . . . . . . . . 57
4.4.3 Computational Complexity and Performance Analyses . . . . . . . 58
4.5 Numerical Analyses for Cooperative Multiple Users Case . . . . . . . . . . . . 58
4.5.1 Analyses on Simulated Signals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
4.5.2 Analyses on Real-World Signals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5 Secure Compressive Spectrum Sensing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
5.1.1 Related Work. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5.1.2 Motivations and Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
5.2 System Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
5.2.1 Networks Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
5.2.2 Signal Processing Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
5.3 Malicious User Detection Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.3.1 Proposed Malicious User Detection Algorithm . . . . . . . . . . . . . . . . 72
5.3.2 Rank Order Estimation Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.3.3 Malicious User Number Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
5.3.4 Analyses on Minimal Number of Active Secondary Users . . . . 79
5.4 Numerical Analyses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
5.4.1 Numerical Results Using Simulated Signals . . . . . . . . . . . . . . . . . . . 81
5.4.2 Numerical Results Using Real-World Signals. . . . . . . . . . . . . . . . . . 85
Contents xv

5.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

Part III Conclusions

6 Conclusions and Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
6.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
6.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Acronyms and Nomenclature

ADC Analog-to-digital conversion

AOP Adaptive outlier pursuit
AWGN Additive white Gaussian noise
CR Cognitive radio
CRN Cognitive radio network
CS Compressive sensing/compressed sensing
CSS Cooperative spectrum sensing
DNRLS Data-assisted non-iteratively reweighted least squares
DSA Dynamic spectrum access
DSO Digital switch-over
DTT Digital terrestrial television
DVB-T Digital Video Broadcasting-Terrestrial
FC Fusion center
FCC Federal Communications Commission
FFT Fast Fourier transform
IDFT Inverse digital Fourier transform
i.i.d. Independent and identically distributed
IRLS Iteratively reweighted least squares
MC Matrix completion
NGR National grid reference
NOMA Non-orthogonal multiple access
Ofcom Office of Communications
OFDM Orthogonal frequency division multiplexed
OMP Orthogonal matching pursuit
PMSE Programme making and special events
PU Primary user
QMUL Queen Mary University of London
RF Radio frequency
RIP Restricted isometry property
ROC Receiver operating characteristics
RTRMC Riemannian Trust-Region for MC

xvii
xviii Acronyms and Nomenclature

SNR Signal-to-noise ratio

SU Secondary user
TVWS TV white space
UHF Ultra-high frequency
WSD White space device
d Decision on spectrum occupancy
F −1 Inverse discrete Fourier transform
h (t) Channel coefficients in time domain
hf Channel coefficients in frequency domain
Hf Channel coefficients in frequency domain in matrix format
I Number of channels among the spectrum of interest
Imax Maximal iteration number
IN N × N identity matrix
J Number of SUs sensing the same channel at different locations
K Number of occupied channels, or sparsity level, or rank order of a matrix
Kmax Statistical upper bound of the rank order K
K̂ Estimated sparsity level K
L Number of channels in a channel group
Lc Number of corrupted channels
L̂c Estimated number of corrupted channels
M Low-rank matrix
M̂ Recovered low-rank matrix
N Number of samples at Nyquist rates
p Norm
pij Uncorrupted power value of the ith channel sensed by the j th SU
p̃ij Corrupted power value of the ith channel sensed by the j th SU
p̂ij Recovered power value of the ith channel sensed by the j th SU
P Number of compressed measurements
Pas Average received power of the wanted DTT signal
Pd Probability of detection
P̄d Target probability of detection
Pf Probability of false alarm
P̄f Target probability of false alarm
PI B Maximum allowable EIRP
PΩ Complete matrix constructed at the FC
r (t) Received signal in time domain
rf Received signal in frequency domain
Rf Received signal in frequency domain in matrix format
s (t) Transmitted signal in time domain
sf Transmitted signal in frequency domain
ŝf Recovered signal
Sf Transmitted signal in frequency domain in matrix format
Ŝf Recovered signal in matrix format
w (t) Additive white Gaussian noise in time domain
wf AWGN in frequency domain
Acronyms and Nomenclature xix

Wf AWGN in frequency domain in matrix format

W Weights matrix for IRLS
W̃ Newly constructed weights matrix for DNRLS
x Compressed measurements
X Compressed measurements in matrix format
γ Compression ratio
κ Malicious user ratio
λ Threshold for energy detection
Λ Binary matrix denoting the uncorrupted channels
Ω Index set of the complete matrix
Φ Measurement matrix
Ψ Sparsifying matrix
ε Error tolerance
 Part I
Background
Chapter 1
Introduction

Radio frequency (RF) spectrum is a valuable but tightly regulated resource due
to its unique and important role in wireless communications. The demand for
RF spectrum is increasing due to a rapidly expanding market of multimedia
wireless services, while the usable spectrum is becoming scarce due to current
rigid spectrum allocation policies. Specifically, according to reports from the
Federal Communications Commission (FCC) and the Office of Communications
(Ofcom), localized temporal and geographic spectrum utilization is extremely low
in reality (Kolodzy and Avoidance 2002; Ofc a). Cognitive radio (CR) has become
a promising solution to solve the spectrum scarcity problem, by allowing secondary
users (SUs) to opportunistically access a licensed band when the primary user
(PU) is absent (Mitola and Maguire 1999). Additionally, it is demonstrated that TV
spectrum, which is used to be allocated to analog TV signals, has been cleaned and
opened to access due to the digital switch-over (DSO) around the world (Kolodzy
and Avoidance 2002; Ofc a). These underutilized TV spectra are named as TV white
space (TVWS). Lately, FCC issued a report and order for permitting the cognitive
usage of TVWS spectrum (fcc). Most recently, Ofcom has enabled license exempt
use of TVWS to harness the benefits of such an innovative wireless technology (Ofc
b), which motivates the further research on the cognitive access to TVWS spectrum.
In order to avoid any harmful interference to the PUs in TVWS, SUs in CR
networks (CRNs) should be aware of the spectrum occupancy over TV band.
Spectrum sensing is the first and one of the most challenging tasks in CR, which is
performed to detect the spectrum holes over TV spectrum. As the radio environment
changes over time and space, an efficient spectrum sensing technique should be
capable of tracking these fast changes (Akyildiz et al. 2006). A good approach for
detecting the primary transmitters is to adopt the traditional narrowband sensing
algorithms, which include energy detection, matched-filtering, and cyclostationary
feature detection. Here, the term “narrowband” implies that the frequency range is
sufficiently narrow, such that the channel frequency response can be considered as

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2019 3

Y. Gao, Z. Qin, Data-Driven Wireless Networks, SpringerBriefs in Electrical
and Computer Engineering, https://doi.org/10.1007/978-3-030-00290-9_1
4 1 Introduction

flat. In another word, the bandwidth of interest is less than the coherence bandwidth
of the channel (Sun et al. 2013).
While the present spectrum sensing algorithms have focused on exploiting
spectral opportunities over narrow frequency range, CRNs will eventually be
required to exploit spectral opportunities over wide frequency range from hundreds
of megahertz (MHz) to several gigahertz (GHz), in order to improve spectrum
efficiency and achieve higher opportunistic throughput. In wideband spectrum
sensing, as driven by the Nyquist sampling theory, a simple approach is to acquire
the wideband signal directly by a high-speed analog-to-digital converter (ADC).
So far, wideband spectrum sensing has been investigated in Tian and Giannakis
(2007), Farhang-Boroujeny (2008), Quan et al. (2009), Sun et al. (2010) with the
implementation of a high-speed ADC. However, the high-speed ADC is particularly
challenging or even unaffordable for energy-constrained devices, such as smart
phones or even battery-free devices in a wireless power transfer model (Qin et al.
2017). Subsequently, Landau (1967) demonstrated that sampling rate should be
no less than the measure of occupied part of the spectrum, with the purpose of
guaranteeing the stable reconstruction of multiband signals. However, the energy
consumption is still unaffordable for energy-constrained SUs in CRNs. Therefore,
revolutionary wideband spectrum sensing techniques become more than desired to
release the burden on high-speed ADCs.
Recent developments on compressive sensing (CS) theory inspire sub-Nyquist
sampling, by utilizing the sparse nature of signals (Candes 2006). Driven by the
inborn nature of the signal sparsity in wireless communications, e.g., the sparse
utilization of spectrum, CS theory is capable of enabling sub-Nyquist sampling
possible for wideband spectrum sensing. More particularly, CS theory has been
firstly applied to wideband spectrum sensing by Tian and Giannakis (2007), where
fewer compressed measurements are required on the basis of Nyquist sampling
theory. Subsequently, the application of CS theory on wideband spectrum sensing
in CRNs has attracted much attention.

1.1 Motivations and Contributions

Along with the developments on CS theory, this book spans the sub-Nyquist-
based wideband spectrum sensing with particular emphasis on CS technique.
These proposed algorithms are capable of improving the robustness and security
of CRNs, with low computational complexity at energy-constrained SUs. The
specific motivations and contributions of this book research are summarized in the
following.
1.1 Motivations and Contributions 5

1.1.1 Data-Driven Compressive Spectrum Sensing

Besides the robustness to channel noise, adaptive compressive spectrum sensing

with low complexity has attracted much attention (Zhang et al. 2018). Theoretically,
the required number of measurements will proportionally change when the sparsity
level of wideband signal varies. However, in practice, the sparsity level of wideband
signal is uncertainty, because of either the dynamic activities of PUs or the
time-varying fading channels between PUs and SUs. Consequently, most of sub-
Nyquist wideband sensing systems should pessimistically choose the number of
measurements to ensure exact recovery, leading to more energy consumption at SUs.
Moreover, the computational complexity of signal recovery may be unaffordable
for the energy-constrained SUs as it is dependent on the number of collected
compressed measurements. Therefore, a low-complexity compressive spectrum
sensing algorithm is needed, which should be adaptive to the dynamic spectrum
occupancy.
Inspired by the geolocation database for TVWS, which is another approach to
make SUs aware of spectrum occupancy, a hybrid framework combining compres-
sive spectrum sensing and geolocation database is proposed to achieve adaptive CS
with low complexity (Qin et al. 2015, 2016a; Gao et al. 2016). More specifically,
a geolocation database algorithm is proposed to be implemented at SUs locally
to provide prior information on the spectrum occupancy. As a result, SUs collect
samples at the minimum rate without loss of any information. Additionally, with
the availability of prior information, a data-assisted non-iteratively reweighted least
squares (DNRLS)-based compressive spectrum sensing algorithm is proposed to
reduce the computational complexity of signal recovery. In order to further improve
accuracy and efficiency of the geolocation database algorithm implemented at SUs,
an efficient approach for calculating the maximum allowable equivalent isotropic
radiated power (EIRP) is proposed. Furthermore, the proposed hybrid framework
and algorithms are tested on the real-world signal and data over TVWS after being
approved by the simulated data.

1.1.2 Robust Compressive Spectrum Sensing

With the use of CS at SUs, each SU would only collect compressed samples at
sub-Nyquist sampling rate. Subsequently, signal recovery would be performed at
SUs or a fusion center (FC), where the data from the spatially located SUs are
fused. It is noticed that the signal-to-noise ratio (SNR) of the CS measurements
would be decreased by 3 dB for every octave increasing in the subsampling factor
for acquisition of a noisy signal with fixed sparsity level (Treichler et al. 2009). This
makes the exact signal recovery more difficult for compressive spectrum sensing
under heavy channel noise. Therefore, a robust spectrum sensing algorithm based
on CS with low computational complexity is needed.
6 1 Introduction

As motivated by this, two robust compressive spectrum sensing algorithms are

designed for the single SU case and the case with multiple SUs, respectively (Qin
et al. 2014, 2016b). The proposed algorithms contain two phases. In the case
with single SU, where signal recovery is to be performed at the SU locally, a
new wideband channel division scheme is proposed to reduce the computational
complexity of signal recovery in the first phase. In the second phase, a denoising
algorithm is performed to improve detection performance by enabling the com-
pressive spectrum sensing algorithm being more robust to channel noise. For the
case with multiple cooperative SUs, where spatial diversity among participating
SUs is utilized to improve the sensing performance (Ghasemi and Sousa 2005;
Akyildiz et al. 2011), the sparse property of spectral signals can be transformed into
a low-rank property (Wang et al. 2012). In the first phase, the proposed wideband
channel division scheme is invoked to reduce the costs of signal acquisition at SUs.
Subsequently, only the compressed measurements are sent to the FC, which reduces
the amount of transmission overhead in CRNs. Matrix completion (MC), as a further
development of CS, is invoked at the FC to recover the unsensed channels from the
sensed channels. In the second phase, detection performance is further improved
by the proposed denoising algorithm. To this end, the proposed robust compressive
spectrum sensing algorithm is tested on the real-world signals over TVWS after
being validated by the simulated TV signals.

1.1.3 Secure Compressive Spectrum Sensing

Along with improving the robustness, adaption, and reducing the complexity
of compressive spectrum sensing algorithm, another challenge for CRNs comes
from the malicious users, which will send out dishonest data to degrade system
performance. In current CSS networks, all cooperative SUs are assumed to be honest
and genuine. However, the existence of malicious users would severely degrade the
performance of cooperative spectrum sensing (CSS) networks. Moreover, malicious
users can degrade the detection performance heavily in sub-Nyquist-based CSS
networks. If part of the compressed measurements are corrupted by malicious users,
signal recovery would be unstable at the FC.
In order to guarantee the security of CSS networks, a malicious user detection
framework is proposed by invoking the low-rank MC technique (Zhang et al. 2014;
Qin et al. 2018). More specifically, with the purpose of improving the detection
accuracy and reducing the costs of data acquisitions at SUs, the data corrupted by
malicious users are removed during the MC process at the FC. Additionally, in
order to avoid requiring any prior information of the CSS networks, a rank order
estimation algorithm and a malicious user number estimation strategy are proposed.
The proposed framework is tested on the real-world signals over TVWS after being
validated by the simulated TV signals. Numerical results show that the proposed
malicious user detection framework achieves higher detection accuracy with lower
costs of data acquisition at SUs or less number of active SUs.
References 7

References

Akyildiz, I. F., Lee, W.-Y., Vuran, M. C., & Mohanty, S. (2006). Next generation/dynamic spectrum
access/cognitive radio wireless networks: A survey. Computer Network, 50, 2127–2159.
Akyildiz, I. F., Lo, B. F., & Balakrishnan, R. (2011). Cooperative spectrum sensing in cognitive
radio networks: A survey. Physical Communication, 4, 40–62.
Candes, E. (2006). Compressive sampling. In Proceedings of the International Congress of
Mathematicians, Madrid, Spain (vol. 3, pp. 1433–1452)
Farhang-Boroujeny, B. (2008). Filter bank spectrum sensing for cognitive radios. IEEE Transac-
tions on Signal Processing, 56, 1801–1811.
Federal Communications Commission (FCC). (2008). Second report and order and memorandum
opinion and order in matter of unlicensed operation in the TV broadcast bands, additional
spectrum for unlicensed devices below 900 MHz and in the 3 GHz band, Document 08-260.
Gao, Y., Qin, Z., Feng, Z., Zhang, Q., Holland, O., & Dohler, M. (2016). Scalable and reliable
IoT enabled by dynamic spectrum management for M2M in LTE-A. IEEE Internet of Things
Journal, 3, 1135–1145.
Ghasemi, A., & Sousa, E. (2005). Collaborative spectrum sensing for opportunistic access in fading
environments. In Proceedings of the IEEE International Symposium on Dynamic Spectrum
Access Networks (DYSPAN), Baltimore, MD (pp. 131–136)
Kolodzy, P., & Avoidance, I. (2002). Spectrum policy task force. Federal Communications
Commission, Washington, DC, Rep. ET Docket.
Landau, H. (1967). Necessary density conditions for sampling and interpolation of certain entire
functions. Acta Mathematica, 117, 37–52.
Mitola, J., & Maguire, G. Q. (1999). Cognitive radio: Making software radios more personal. IEEE
Personal Communications, 6, 13–18.
Qin, Z., Gao, Y., & Parini, C. G. (2016a). Data-assisted low complexity compressive spectrum
sensing on real-time signals under sub-Nyquist rate. IEEE Transactions on Wireless Communi-
cations, 15, 1174–1185.
Qin, Z., Gao, Y., Plumbley, M., & Parini, C. (2014). Efficient compressive spectrum sensing
algorithm for M2M devices. In IEEE Global Conference on Signal and Information Processing
(GlobalSIP), Atlanta, GA (pp. 1170–1174).
Qin, Z., Gao, Y., & Plumbley, M. D. (2018). Malicious user detection based on low-rank matrix
completion in wideband spectrum sensing. IEEE Transactions on Signal Processing, 66, 5–17.
Qin, Z., Gao, Y., Plumbley, M. D., & Parini, C. G. (2016b). Wideband spectrum sensing on real-
time signals at sub-Nyquist sampling rates in single and cooperative multiple nodes. IEEE
Transactions on Signal Processing, 64, 3106–3117.
Qin, Z., Liu, Y., Gao, Y., Elkashlan, M., & Nallanathan, A. (2017). Wireless powered cognitive
radio networks with compressive sensing and matrix completion. IEEE Transactions on
Communications, 65, 1464–1476.
Qin, Z., Wei, L., Gao, Y., & Parini, C. (2015). Compressive spectrum sensing augmented by geo-
location database. In Proceedings of the International Workshop on Smart Spectrum at IEEE
Wireless Communications and Networking Conference (WCNC), New Orleans, LA (pp. 170–
175).
Quan, Z., Cui, S., Sayed, A. H., & Poor, H. V. (2009). Optimal multiband joint detection for
spectrum sensing in cognitive radio networks. IEEE Transactions on Signal Processing, 57,
1128–1140.
Sun, H., Laurenson, D. I., & Wang, C. X. (2010). Computationally tractable model of energy
detection performance over slow fading channels. IEEE Communications Letters, 14, 924–926.
Sun, H., Nallanathan, A., Wang, C.-X., & Chen, Y. (2013). Wideband spectrum sensing for
cognitive radio networks: A survey. IEEE Wireless Communications, 20, 74–81.
Tian, Z., & Giannakis, G. (2007). Compressed sensing for wideband cognitive radios. In IEEE
International Conference on Acoustics, Speech, and Signal Processing, Honolulu, HI (ICASSP)
(pp. 1357–1360).
8 1 Introduction

Treichler, J., Davenport, M., & Baraniuk, R. (2009). Application of compressive sensing to the
design of wideband signal acquisition receivers. In US/Australia Joint Work. Defense Apps of
Signal Processing (DASP) (vol. 5).
UK Office of Communications (Ofcom). (2009). Statement on cognitive access to interleaved
spectrum.
UK Office of Communications (Ofcom). (2015). Decision to make the wireless telegraphy (White
Space Devices).
Wang, Y., Tian, Z., & Feng, C. (2012). Collecting detection diversity and complexity gains in
cooperative spectrum sensing. IEEE Wireless Communications, 11, 2876–2883.
Zhang, X., Ma, Y., Gao, Y., & Zhang, W. (2018). Autonomous compressive sensing augmented
spectrum sensing. IEEE Transactions on Vehicular Technology, 67, 6970–6980.
Zhang, X., Qin, Z., & Gao, Y. (2014). Dynamic adjustment of sparsity upper bound in wideband
compressive spectrum sensing. In Proceedings of the IEEE Global Conference on Signal and
Information Processing (GlobalSIP), Atlanta, GA (pp. 1214–1218).
Chapter 2
Sparse Representation in Wireless
Networks

Sparse representation of signals has received extensive attention due to its capacity
for efficient signal modeling and related applications. The problem solved by the
sparse representation is to search for the most compact representation of a signal
in terms of a linear combination of the atoms in an overcomplete dictionary. In
the literature, three aspects of research on the sparse representation have been
focused:
1. Pursuit methods for solving the optimization problem, such as matching pursuit
and basis pursuit;
2. Design of the dictionary, such as the K-SVD method;
3. Applications of the sparse representation, such as wideband spectrum sensing,
channel estimation of massive MIMO, and data collection in WSNs.
General sparse representation methods, such as principal component analysis
(PCA) and independent component analysis (ICA), aim to obtain a representation
that enables sufficient reconstruction. It has been demonstrated that PCA and ICA
are able to deal with signal corruption, such as noise, missing data, and outliers. For
sparse signals without measurement noise, CS can recover the sparse signals exactly
with random measurements. Furthermore, the random measurements significantly
outperform measurements based on PCA and ICA for the sparse signals without
corruption (Chang et al. 2009b; Wright et al. 2010; Chang et al. 2009a).

2.1 Principles of Standard Compressive Sensing

The principles of standard CS, such as to be performed at a single node, can be

summarized in the following three parts (Candes 2006).

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2019 9

Y. Gao, Z. Qin, Data-Driven Wireless Networks, SpringerBriefs in Electrical
and Computer Engineering, https://doi.org/10.1007/978-3-030-00290-9_2
10 2 Sparse Representation in Wireless Networks

2.1.1 Sparse Representation

Generally speaking, sparse signals contain much less information than their ambient
dimension suggests. Sparsity of a signal is defined as the number of non-zero
elements in the signal under a certain domain. Let f be an N -dimensional signal
of interest, which is sparse over the orthonormal transformation basis matrix Ψ ∈
RN ×N , and s be the sparse representation of f over the basis Ψ . Then f can be given
by

f = Ψ s. (2.1)

Apparently, f can be the time or space domain representation of a signal, and s is the
equivalent representation of f in the Ψ domain. For example, if Ψ is the inverse
Fourier transform (FT) matrix, then s can be regarded as the frequency domain
representation of the time domain signal, f. Signal f is said to be K-sparse in the Ψ
domain if there are only K (K  N ) out of the N coefficients in s that are non-zero.
If a signal is able to be sparsely represented in a certain domain, the CS technique
can be invoked to take only a few linear and non-adaptive measurements.

2.1.2 Projection

When the original signal f arrives at the receiver, it is processed by the measurement
matrix Φ ∈ RP ×N with P < N, to get the compressed version of the signal, that is,

x = Φf = ΦΨ s=Θs, (2.2)

where Θ = ΦΨ is a P × N matrix, called the sensing matrix. As Φ is independent

of signal f, the projection process is non-adaptive.
Figure 2.1 illustrates how the different sensing matrices Θ influence the projec-
tion of a signal from high dimension
 to its space, i.e., mapping s ∈ R3 to x ∈ R2 .
As shown in Fig. 2.1, s = s s 0 is a three-dimensional signal. When s is mapped
 
1 −1 0
into a two-dimensional space by taking Θ1 = as the sensing matrix,
0 0 1
the original signal s cannot be recorded based on the projection under Θ1 . This is
because that the plane spanned by the two row vectors of Θ1 is orthogonal to signal s
as shown in Fig. 2.1a. Therefore, Θ1 corresponds to the worst projection.  As shown

100
in Fig. 2.1b, we can also observe that the projection by taking Θ2 = is
001
not a good one. It is noted that the plane spanned by the two row vectors of Θ2 can
only contain part of information of the sparse signal s, and the sparse component in
the direction of s2 is missed when the signal s is projected into the two-dimensional
2.1 Principles of Standard Compressive Sensing 11

(a) (b) (c)

s2 s2 s2

è è
æs ç æs ç
æs
è
ç
S = çs S = çç s
ç ç
ç ç ç
ç ç S =çs ç
ç0 æ ç0 æ
ç0
è è è
æ

s1 s1 s1

s3 s3 s3
è è è
æ1 −1 0 ÷ æ1 0 0 ÷ æ1 1 0 ÷
Θ1 = ÷ æ Θ 2 =÷ æ Θ3 = ÷ æ
è0 0 1 è0 0 1 è0 0 1

Fig. 2.1 Projection of a sparse signal with one non-zero component with different sensing
matrices Qin et al. (2018). (a) Worst projection. (b) Bad projection. (c) Good projection

 
110
space. When the sensing matrix is set to Θ3 = , as shown in Fig. 2.1c, the
001
signal s can be fully recorded as it falls into the plane spanned by the two row vectors
of Θ3 . Therefore, Θ3 results in a good projection and s can be exactly recovered by
its projection x in the two-dimensional space. Then it is natural to ask what type of
projection is good enough to guarantee the exact signal recovery?
The key of CS theory is to find out a stable basis Ψ or measurement matrix Φ to
achieve exact recovery of the signal with length N from P measurements. It seems
an undetermined problem as P < N. However, it has been proved in Candes et al.
(2006) that exact recovery can be guaranteed under the following conditions:
• Restricted isometry property (RIP): Measurement matrix Φ has the RIP of order
K if

Φf22
1 − δK ≤ ≤ 1 + δK (2.3)
f22

holds for all K-sparse signal f, where δK is the restricted isometry constant of a
matrix Φ.
• Incoherence property: Incoherence property requires that the rows of measure-
ment matrix Φ cannot sparsely represent the columns of the sparsifying matrix
Ψ and vice versa. More specifically, a good measurement will pick up a little bit
information of each component in s based on the condition that Φ is incoherent
with Ψ . As a result, the extracted information can be maximized by using the
minimal number of measurements.
It has been pointed out that verifying both the RIP condition and incoherence
property is computationally complicated but they could be achieved with a high
12 2 Sparse Representation in Wireless Networks

probability simply by selecting Φ as a random matrix. The common random

matrices include Gaussian matrix, Bernoulli matrix, or almost all others matrices
with independent and identically distributed (i.i.d.) entries. Besides, with the
properties of the matrix with i.i.d. entries Φ, the matrix Θ = ΦΨ is also random
i.i.d., regardless of the choice of Ψ . Therefore, the random matrices are universal
as they are random enough to be incoherent with any fixed basis. It has been
demonstrated that random measurements can universally capture the information
relevant for many compressive signal processing applications without any prior
knowledge of either the signal class and its sparse domain or the ultimate signal
processing task.
Moreover, for Gaussian matrices the number of measurements required to
guarantee the exact signal recovery is almost minimal. However, random matrices
inherently have two major drawbacks in practical applications: huge memory
buffering for storage of matrix elements, and high computational complexity due
to their completely unstructured nature (Candes and Romberg 2007). Compared to
the standard CS that limits its scope to standard discrete-to-discrete measurement
architectures using random measurement matrices and signal models based on
standard sparsity, more structured sensing architectures, named structured CS,
have been proposed to implement CS on feasible acquisition hardware. So far,
many efforts have been put on the design of structured CS matrices, i.e., random
demodulator (Tropp et al. 2010), to make CS implementable with expense of
performance degradation. Particularly, the main principle of random demodulator
is to multiply the input signal with a high-rate pseudonoise sequence, which
spreads the signal across the entire spectrum. Then a low-pass anti-aliasing filter
is applied and the signal is captured by sampling it at a relatively low rate. With
the additional digital processing to reduce the burden on the analog hardware,
random demodulator bypasses the need for a high-rate analog-to-digital converter
(ADC) (Tropp et al. 2010). A comparison of Gaussian sampling matrix and random
demodulator is provided in Fig. 2.2 in terms of detection probability with different
compression ratios P /N. From the figure, the Gaussian sampling matrix performs
better than the random demodulator.

2.1.3 Signal Reconstruction

After the compressed measurements are collected, the original signal should be
reconstructed. Since most of the basis coefficients in s are negligible, the original
signal can be reconstructed by finding out the minimal set of coefficients that
matches the set of compressed measurements x, that is, by solving

ŝ = arg min sp subject to Θs = x, (2.4)

s
2.2 Reweighted Compressive Sensing 13

0.8

0.6
Pd

0.4

0.2
Random demodulator
Gaussian distributed matrix
0
10-2 10-1 100
P/N

Fig. 2.2 Detection probability versus compression ratio with different measurement matrices. In
this case, the signal is one-sparse

where ·p is the p -norm and p = 0 corresponds to counting the number of non-
zero elements in s. However, the reconstruction problem in (2.4) is both numerically
unstable and NP-hard (Candes 2006) when 0 -norm is used.
So far, there are mainly two types of relaxations to problem (2.4) to find a
sparse solution. The first type is convex relaxation, where 1 -norm is used to
substitute 0 -norm in (2.4). Then (2.4) can be solved by standard convex solvers,
e.g., cvx. It has been proved that 1 norm results √
in the same solution as 0 norm
when RIP is satisfied with the constant δ2k < 2 − 1 (Cands 2008). Another
type of solution is to use a greedy algorithm, such as OMP (Tropp and Gilbert
2007), to find a local optimum in each iteration. In comparison with the convex
relaxation, the greedy algorithm usually requires lower computational complexity
and time cost, which makes it more practical for wireless communication systems.
Furthermore, the recent result has shown that the recovery accuracy achieved by
some greedy algorithms is comparable to the convex relaxation but requiring much
lower computational cost (Choi et al. 2017).

2.2 Reweighted Compressive Sensing

As aforementioned, 1 -norm is a good approximation for the NP-hard 0 -norm

problem when RIP holds. However, the large coefficients are penalized more
heavily than the small ones in 1 -norm minimization, which leads to performance
degradation on signal recovery. To balance the penalty on the large and the
small coefficients, reweighted CS is introduced by providing different penalties on
those large and small coefficients. A reweighted 1 -norm minimization framework
14 2 Sparse Representation in Wireless Networks

(Candes 2006) has been developed to enhance the signal recovery performance with
fewer compressed measurements by solving

ŝ = arg min Ws1 subject to Θs = x, (2.5)

s

where W is a diagonal matrix with w1 , . . . , wn on the diagonal and zeros elsewhere.

Moreover, p -norm, e.g., 0 < p < 1, is utilized to lower the computational
complexity of signal recovery process caused by the 1 -norm optimization problem.
Iterative reweighted least square (IRLS)-based CS approach has been proposed
in Rao and Kreutz-Delgado (1999) to solve (2.4) in a non-convex approach as


N
ŝ = arg min wi si subject to Θs = x, (2.6)
s
i=1

 
 (l−1) p−2 (l−1)
where wi = si  is computed based on the result of the last iteration, si .
It is worth noting that (2.4) becomes non-convex when p < 1. The existing
algorithms cannot guarantee to reach a global optimum and may only produce local
minima. However, it has been proved (Chartrand 2007; Chartrand and Staneva 2008)
that under some circumstances the reconstruction in (2.4) will reach a unique and
global minimizer (Chartrand and Yin 2008), which is exactly ŝ = s. Therefore, we
can still exactly recover the signal in practice.

2.3 Distributed Compressive Sensing

The distributed compressive sensing (DCS) (Baron et al. 2009) is an extension

of the standard one by considering networks with M nodes. At the m-th node,
measurement xm can be given by

xm = Θm sm , ∀m ∈ M , (2.7)

where M is the set of nodes in the network. As stated in (2.2), Θm is the sensing
matrix deployed at the m-th node, and sm is a sparse signal of interest. DCS becomes
a standard CS when M = 1.
In the applications of standard CS, the signal received at the same node has its
sparsity property due to its intra-correlation. While for the networks with multiple
nodes, signals received at different nodes exhibit strong inter-correlation. The intra-
correlation and inter-correlation of signals from the multiple nodes lead to a joint
sparsity property. The joint sparsity level is usually smaller than the aggregate
over the individual signal’s sparsity level. As a result, the number of compressed
measurements required for exact recovery in DCS can be reduced significantly
compared to the case performing standard CS at each single node independently.
2.4 Compressive Spectrum Sensing 15

In DCS, there are two closely related concepts: distributed networks and
distributed CS solvers. The distributed networks refer to networks that different
nodes perform data acquisition in a distributed way and the standard CS can
be applied at each node individually to perform signal recovery. While for DCS
solver as proposed in Baron et al. (2009), the data acquisition process requires no
collaboration among sensors and the signal recovery process is performed at several
computational nodes, which can be distributed in a network or locally placed within
a multiple core processor. Generally, it is of interest to minimize both computation
cost and communication overhead in DCS. The most popular application scenario of
DCS is that all signals share the common sparse support but with different non-zero
coefficients.

2.4 Compressive Spectrum Sensing

The last decade has witnessed the rapid explosion of wireless devices all over the
world, which gives rise to the increasing demand for wireless spectral resource.
As reported by FCC and Ofcom (Kolodzy and Avoidance 2002; Ofc a), there are
significant temporal and spatial variations in the allocated spectrum. Given this
fact, CR has been proposed as an intelligent system to detect spectrum holes for
unlicensed usage (Mitola and Maguire 1999). More specifically, the basic idea of CR
is to match the requirements of higher layer applications or users with the available
resources. The available resources include available power, spectrum, and other
resources that can be utilized by unlicensed SUs. CR is a radio that is capable of
sensing the available resources and learning from the user behaviors and its previous
decisions and mistakes, in order to provide a better response to the new resource
request from SUs. So far, CR has been widely investigated.

2.4.1 Spectrum Sensing Methods

In CR, spectrum sensing is one of the most challenging tasks, which allows SUs to
have the knowledge of spectrum occupancy. Once a spectrum hole is detected, SUs
can make use of it for data transmission. Spectrum sensing requires high accuracy
and low complexity for DSA (Nekovee 2008). There is an extensive research work
on spectrum sensing techniques being carried out. Many theoretical models for
spectrum sensing techniques have been proposed, such as matched filter detection,
cyclostationary feature detection, and energy detection. The matched filter detection
is an optimal detection method that requires the prior information of PUs (Bhargavi
and Murthy 2010). However, it requires SUs to have a dedicated sensing receiver
for each type of PU signals. Cyclostationary feature detection can distinguish the
PUs and noise by utilizing the periodicity in the received primary signal. However,
it requires high computational complexity and prior information of the primary
Exploring the Variety of Random
Documents with Different Content
Dowland shook his head. He could think about that when he had
more time. He'd lost—he looked at his watch—a little less than four
hours. In four hours, a large number of things might have happened
in the ranch area, with only the one partly attractive possibility
among them that somebody had managed to get into the laboratory
and shut off the YM flow.

He started back at a cautious trot. Downhill and with the light

strengthening gradually, covering ground was considerably less of a
problem than it had been during the night. The wind hadn't let up; it
still came in wild, intermittent gusts that bent the trees. Now and
then a cloud of dust whipped past, suggesting that the air over the
desert was also violently disturbed. And it might very well be,
Dowland thought, that YM could upset atmospheric conditions in an
area where it was active. Otherwise, if there was anything abnormal
going on in the forest about him, there were no detectable
indications of it.
He came out presently on a ridge from where the ranch area was in
view. It lay now approximately a third of a mile ahead. In the dim
light, everything seemed quiet. Dowland slowed to a walk.
He might be heading into an ambush down there. Jill Trelawney
could, at most, be beginning to wake up from her drugged sleep and
for another hour or so she would be too confused and groggy to
present a problem. But others might be at the ranch by now; Paul
Trelawney or a group of Carter's Troopers. And whether Jill was able
to give them a coherent report or not, any of the Freeholder
conspirators would discover very quickly that somebody who was not
a member of their group had been there before them; they would
anticipate his return, be on the watch for it. Dowland left the direct
line he had been following, and headed east, moving with constantly
increasing caution. On that side, the forest grew closest to the ranch
buildings, and he remembered noticing a hedge-like thicket of
evergreens just north of the cleared land. He could make a
preliminary check of the area from there.
He was within a hundred and fifty feet of the point when he
discovered just how healthy the notion of a preliminary check had
been. A man was lying in the cover of the evergreens Dowland had
been thinking about, head up, studying the ranch grounds. He wore
an antiradiation suit of the type Dowland had found in the storeroom;
a heavy rifle lay beside him. His face was in profile. It was smeared
now with the sweat and dirt the AR field had held in, but Dowland
recognized the bold, bony features instantly.
He had finally found Doctor Paul Trelawney.

It took Dowland over eight minutes to cover the remaining distance

between them. But the stalk had eminently satisfactory results. He
was within a yard of Trelawney before the Freeholder became aware
of his presence. The IPA gun prodded the man's spine an instant
later.
"No noise, please," Dowland said softly. "I'd sooner not kill you. I
might have to."
Paul Trelawney was silent for a moment. When he spoke, his voice
was raw with shock. "Who the devil are you?"
"Solar Police Authority," Dowland said. "You know why I'm here."
Trelawney grunted. Dowland went on, "Why are you hiding out?"
"Why do you think?" Trelawney asked irritably. "Before showing
myself, I was trying to determine the whereabouts of the man who
fired a rifle within half a mile of me during the night."
So they had been stalking each other. Dowland said, "Why couldn't
that person have been your brother or niece?"
"Because I know the sound of our rifles."
"My mistake.... Do you have a gun or other weapon on you?"
"A knife."
"Let's have it."
Trelawney reached under his chest, brought out a sheathed knife and
handed it back to Dowland. Dowland lobbed it into the bushes a few
yards away, moved back a little.
"Get up on your hands and knees now," he said, "and we'll make sure
that's all."
He was careful about the search. Trelawney appeared passive enough
at the moment, but he was not a man too take chances with. The AR
suit turned out to be concealing a tailored-in two-way communicator
along with as many testing and checking devices as an asteroid
miner's outfit, but no weapons. In a sealed pocket, obviously
designed for it, was a five-inch atomic key. Dowland slid the heavy
disk out with fingers that suddenly were shaking a little.
"Does this open your laboratory here?"
"Yes."
Dowland detached the communicator's transmission unit, and
dropped it with the laboratory key into his pocket. "All right," he said,
"turn around and sit down." He waited until Trelawney was facing
him, then went on. "How long have you been watching the ranch?"
"About an hour."
"Seen anyone—or anything?"
Trelawney regarded him quizzically, shook his head. "Not a thing."
"I won't waste time with too many questions just now," Dowland
said. "The laboratory is locked, and the machine you started in there
apparently is still in operation. Your brother was found outside the
laboratory yesterday morning, and may be dead or dying of internal
radiation burns. He was alive and didn't seem to be doing too badly
when I left him and Miss Trelawney in the house last night to go
looking for you. I had to drug Miss Trelawney—she isn't a very
cooperative person. She should still be asleep.
"Now, if I hadn't showed up here just now, what did you intend to
do?"
"I intended to stop the machine, of course," Trelawney said. His
expression hadn't changed while Dowland was talking. "Preferably
without involving the Solar Police Authority in our activities. But since
you've now involved yourself, I urgently suggest that we go to the
laboratory immediately and take care of the matter together."

Dowland nodded. "That's what I had in mind, Trelawney. Technically

you're under arrest, of course, and you'll do whatever has to be done
in there at gun point. Are we likely to run into any difficulties in the
operation?"
"We very probably will," Trelawney said thoughtfully, "and it's just as
probable that we won't know what they are before we encounter
them."
Dowland stood up. "All right," he said, "let's go. We'll stop off at the
house on the way. I want to be sure that Miss Trelawney isn't in a
position to do something thoughtless."
He emptied the magazine of Trelawney's rifle before giving it to him.
They started down to the house, Trelawney in the lead, the IPA gun
in Dowland's hand.
The house door was closed. Trelawney glanced back questioningly.
Dowland said in a low voice, "It isn't locked. Open it, go on in, and
stop two steps inside the hallway. I'll be behind you. They're both in
the living room."
He followed Trelawney in, reaching back to draw the door shut again.
There was a whisper of sound. Dowland half turned, incredulously
felt something hard jab painfully against his backbone. He stood still.
"Drop your gun, Dowland," Jill Trelawney said behind him. Her voice
was as clear and unslurred as if she had been awake for hours.
Dowland cursed himself silently. She must have come around the
corner of the house the instant they went in.
"My gun's pointing at your uncle's back," he said. "Don't do anything
that might make me nervous, Miss Trelawney."
"Don't try to bluff Jill, friend," Paul Trelawney advised him without
turning his head. There was dry amusement in the man's voice. "No
one's ever been able to do it. And she's quite capable of concluding
that trading an uncle for an SPA spy would still leave Terra ahead at
this stage. But that shouldn't be necessary. Jill?"
"Yes, Paul?"
"Give our policeman a moment to collect his wits. This does put him
in a very embarrassing position, after all. And I can use his help in
the lab."
"I'll give you exactly three seconds, Dowland," Jill said. "And you'd
better believe that is not a bluff. One...."
Dowland dropped his gun.

The two Trelawneys held a brief, whispered conversation in the living

room. Dowland, across the room from them, and under cover of two
guns now, couldn't catch much of it. Jill was in one of the radiation
suits he'd brought in from the storeroom. Miguel was dead. He had
still been unconscious when she woke up, and had stopped breathing
minutes afterwards. Medic had done what it could; in this case it
simply hadn't been enough. Jill, however, had found another use for
it. Dowland thought the possibility mightn't have occurred to anyone
else in similar circumstances; but he still should have thought of it
when he left the house. As she began to struggle up from sleep, she
remembered what Dowland had told her about medic, and somehow
she had managed to inject a full ampule of it into her arm. It had
brought her completely awake within minutes.
The murmured talk ended. The girl looked rather white and
frightened now. Paul Trelawney's face was expressionless as he came
over to Dowland. Jill shoved the gun she had put on Dowland into
her belt, picked up Paul's hunting rifle, held it in her hands, and stood
waiting.
"Here's the procedure, Dowland," Trelawney said. "Jill will go over to
the lab with us, but stay outside on guard. She'll watch...."
"Did you tell her," Dowland interrupted, "to keep an eye out for
something that stands twice as high as this house?"
Trelawney looked at him a moment. "So you ran into it," he said. "I
was wondering. It's very curious that ... well, one thing at a time. I
cautioned her about it, as it happens. Now come over to the table."
Dowland remained standing beside the table, while across from him
Trelawney rapidly sketched out two diagrams on a piece of paper.
The IPA gun lay on the table near Trelawney's right hand. There
might have been an outside chance of reaching it if one could have
discounted Jill's watchfulness. Which, Dowland decided, one couldn't.
And he'd seen her reload the rifle she was holding. He stayed where
he was.
Trelawney shoved the paper across to him.
"Both diagrams represent our machine," he said, "and they should
give you a general idea of what you'll see. This wheel here is at the
far side of the console when we come in the door. The wheel is the
flow regulator—the thing you have to keep in mind. There are scale
markings on it. The major markings have the numbers one to five.
Yesterday morning the regulator was set at five—full flow. Spin the
wheel back to one, and the Ym-400 that's been producing the flow
goes inert. Is that clear?"
Dowland nodded. "Clear enough."
"After that," Trelawney remarked, "we may be able to take things a
little easier."
"What's the quantity you're using in there?"
"No real reason I should tell you that, is there? But I will. The sixty-
eight kilograms the Overgovernment's been grieving about are under
the machine platform. We're using all of it." He grinned briefly,
perhaps at Dowland's expression. "The type of job we had in mind
required quantities in that class. Now, about yourself. We're not
murderers. Jill tells me you can't be bribed—all right. What will
happen, when this thing's settled, is that you'll have an attack of
amnesia. Several months of your life will be permanently lost from
your memory, including, of course, everything connected with this
operation. Otherwise you won't be harmed. Understand?"
"I've heard of such things," Dowland said drily.
It wouldn't, however, be done that way. It was the kind of thing told
a man already as good as dead, to keep him from making a
desperate attempt to save himself. The Freeholders really wouldn't
have much choice. Something had loused up their plans here, and if
Dowland either disappeared or was found suffering from a sudden
bout of amnesia, the IPA would turn its full attention on Terra at
once. If he died, his death could be plausibly arranged to look like an
accident or a killing for personal motives. These people were quite
capable of sacrificing one of their group to back such a story up. And
it would pass. Terra was under no more immediate suspicion than
any other world. Dowland had been on a routine assignment.

There were a few brief preparations. Paul Trelawney checked the

batteries in the radiation suits he and Jill were wearing, then
exchanged his set for that of the spare suit. Dowland left his own AR
field off for the moment. It was at least as adequate as the one
developed by the Trelawneys' suits, and in some respects a much
more practical device. But the suit batteries had an effective life of
twenty-four hours, expending them automatically while the suits were
worn. His field would maintain itself for a minimum of an hour and a
half, a maximum of two hours. In this situation, Dowland wasn't sure
how long he would have to depend on the field. A few more minutes
of assured protection might make a difference.
He saw Trelawney studying the mountaineering rig on the floor; then
he picked up the harness and brought it over to him.
"Here, put it on," he said.
"What for?" Dowland asked, surprised.
Trelawney grinned. "We may have a use for it. You'll find out in a
minute or two."
They left the house by a back entrance. Clouds were banked low on
the eastern horizon now; the first sunlight gleamed pale gold beneath
them. In the west the sky was brown with swirling dust. Jill stopped
twenty yards from the laboratory building and stood on the slope,
rifle in hand, watching the men go on. At the door, Dowland switched
on his AR field. Trelawney tossed the disk-shaped key over to him.
"Know how to use it?"
Dowland nodded.
"All right. After you've snapped it in and it releases again, throw it
back to me. It may be the last one around, and we're not taking it
into the laboratory this time. When the door starts moving down,
step back to the right of it. We'll see what the lab is like before we go
in." Trelawney indicated a thimble-sized instrument on his suit.
"This'll tell whether the place is hot at the moment, and
approximately how hot." He waved the IPA gun in Dowland's
direction. "All right, go ahead."
Dowland fitted the key into the central depression in the door,
pressed down, felt the key snap into position with a sharp twisting
motion of its own, released his pressure on it. An instant later, the
key popped back out into his hand. He tossed it back to Trelawney,
who caught it left-handed and threw it over his head in Jill's direction.
The disk thudded heavily into the grass ten feet from her. The girl
walked over, picked it up, and slid it into one of her suit pockets.
The slab of metasteel which made up the laboratory door began
moving vertically downward. The motion stopped when the door's top
rim was still several inches above the level of the sill.
A low droning came from the little instrument on Trelawney's suit. It
rose and fell irregularly like the buzz of a circling wasp. Mingled with
it was something that might have been the hiss of escaping steam.
That was Dowland's detector confirming. The lab reeked with
radiation.
He glanced over at Trelawney.
"Hot enough," the Freeholder said. "We'll go inside. But stay near the
door for a moment. There's something else I want to find out
about...."

Inside, the laboratory was unpartitioned and largely empty, a great

shell of a building. Only the section to the left of the entrance
appeared to have been used. That section was lighted. The light
arose evenly from the surfaces of the raised machine platform
halfway over to the opposite wall. The platform was square, perhaps
twenty feet along its sides. Dowland recognized the apparatus on it
from Trelawney's diagrams. The central piece was an egg-shaped
casing which appeared to be metasteel. Near its blunt end, partly
concealed, stood the long, narrow instrument console. Behind the
other end of the casing, an extension ramp jutted out above the
platform. At the end of the ramp was a six-foot disk that might have
been quartz, rimless, brightly iridescent. It was tilted to the left,
facing the bank of instruments.
"A rather expensive bit of equipment over there, Dowland,"
Trelawney said. "My brother developed the concept, very nearly in
complete detail, almost twenty-five years ago. But a great deal of
time and thought and work came then before the concept turned into
the operating reality on that platform."
He nodded to the left. "That's Miguel's coat on the floor. I wasn't sure
it would still be here. The atomic key you were searching for so
industriously last night is in one of its pockets. Miguel was standing
just there, with the coat folded over his arm, when I saw him last—
perhaps two or three seconds before I was surprised to discover I
was no longer looking at the instrument controls in our laboratory."
"Where were you?" Dowland asked. "Six hundred thousand years in
the past?"
"The instruments showed a fix on that point in time," Trelawney said.
"But this was, you understand, a preliminary operation. We intended
to make a number of observations. We had not planned a personal
transfer for several more weeks. But in case the test turned out to be
successful beyond our expectations, I was equipped to make the
transfer. That bit of optimistic foresight is why I'm still alive."
What was the man waiting for? Dowland asked, "What actually
happened?"
"A good question, I'd like to know the whole answer myself. What
happened in part was that I suddenly found myself in the air, falling
toward a river. It was night and cloudy, but there was light enough to
show it was a thoroughly inhospitable river.... And now I believe"—his
voice slowed thoughtfully—"I believe I understand why my brother
was found outside the closed door of this building. Over there,
Dowland. What does that look like to you?"
Near the far left of the building, beyond the immediate range of the
light that streamed from the machine stand, a big packing crate
appeared to have been violently—and rather oddly—torn apart. The
larger section of the crate lay near the wall, the smaller one
approximately twenty feet closer to the machine platform. Assorted
items with which it had been packed had spilled out from either
section. But the floor between the two points of wreckage was bare
and unlittered. Except for that, one might have thought the crate had
exploded.

"It wasn't an explosion," Trelawney agreed when Dowland said as

much. He was silent a moment, went on, "In this immediate area,
two space-time frames have become very nearly superimposed.
There is a constant play of stresses now as the two frames attempt
to adjust their dissimilarities. Surrounding our machine we have a
spherical concentration of those stresses, and there are moments
when space here is literally wrenched apart. If one were caught at
such an instant—ah!"
To Dowland it seemed that a crack of bright color had showed briefly
in the floor of the building, between the door and the machine
platform. It flickered, vanished, reappeared at another angle before
his ears had fully registered the fact that it was accompanied by a
curiously chopped-off roar of sound. Like a play of lightning. But this
was....
The air opened out before him, raggedly framing a bright-lit three-
dimensional picture. He was staring down across a foaming river to
the rim of a towering green and yellow forest. The crash of the river
filled the building. Something bulky and black at the far left ... but
the scene was gone—
The interior of the laboratory building lay quiet and unchanged before
them again. Dowland said hoarsely, "How did you know what was
going to happen?"
"I was in a position to spend several hours observing it," Trelawney
said, "from the other side. You see now, I think, that we can put your
mountaineer's kit to some very practical use here."
Dowland glanced across the building. "The walls...."
"Metasteel," Trelawney said, "and thank God for that. The building's
sound; the stresses haven't affected it. We'll have some anchor
points. A clamp piton against that wall, six feet above the console
walk and in line with it, another one against the doorframe here, and
we can rope across."
Dowland saw it, unsnapped his harness, fed the end of the
magnerope through the eye of a piton, and twisted it tight. "Are we
going together?" he asked.
Trelawney shook his head. "You're going, Dowland. Sorry about that,
but this is no time for sporting gestures. The rope doesn't eliminate
the danger. But if you find your feet suddenly dangling over the air of
a very old time, you'll still stay here—I hope. If you don't make it
across, I'll follow. We get two chances to shut Ymir down instead of
one. All right?"
"Since you have the gun, yes," Dowland said. "If I had it, it would be
the other way around."
"Of course," Trelawney agreed. He watched in silence then as
Dowland rammed the threaded piton down the muzzle of the gun,
locked it in position, took aim across the machine platform, and fired.
The piton clamp made a slapping sound against the far wall, froze
against it. Dowland gave the loose end of the rope a few tugs, said,
"Solid," cut the rope, and handed the end to Trelawney.
The Freeholder reached up to set a second piton against the
doorframe, fed a loop of the rope through it, and twisted it tight.
Dowland slipped a set of grappling gloves out of the harness, pulled
one over his right hand, tossed the other to Trelawney. "In case," he
said, "you have to follow. Magnerope gets to be wearing on bare
hands."
Trelawney looked briefly surprised, then grinned. "Thanks," he said.
"Can you do it with one glove?"
"No strain at that distance."
"Too bad you're not a Terran, Dowland. We could have used you."
"I'm satisfied," Dowland said. "Any point in waiting now for another
run of those cracks in space before making the trip?"
Trelawney shook his head. "None at all, I'm afraid. From what I saw,
there's no more regularity in those stress patterns than there is in a
riptide. You see how the rope is jerking right now—you'll get pulled
around pretty savagely, I'd say, even if you don't run into open splits
on the way across."

Dowland was fifteen feet from the door, half running with both hands
on the rope, when something plucked at him. He strained awkwardly
sideways, feet almost lifting from the floor. Abruptly he was released,
went stumbling forward a few steps before the next invisible current
tugged at him, pulling him downward now. It was a very much
stronger pull, and for endless seconds it continued to build up. His
shoulders seemed ready to snap before he suddenly came free again.
The rest of the way to the platform remained almost undisturbed, but
Dowland was trembling with tensions before he reached it; he could
feel the drag of the AR field on his breathing. The steps to the
platform were a dozen feet to his right—too far from the rope.
Dowland put his weight on the rope, swung forward and up, let the
rope go and came down on the narrow walk between instrument
board and machine section. The panels shone with their own light; at
the far end he saw the flow-control wheel Trelawney had indicated, a
red pointer opposite the numeral "5." Dowland took two steps toward
it, grasped the wheel, and spun it down.
The pointer stopped at "1." He heard it click into position there.
Instantly, something slammed him sideways against the console, sent
him staggering along it, and over the low railing at the end of the
platform. The floor seemed to be shuddering as he struck it, and then
to tilt slowly. Dowland rolled over, came up on hands and knees,
facing back toward the platform. Daylight blazed again in the building
behind him, and the roar of a river that rolled through another time
filled his ears. He got to his feet, plunged back toward the whipping
rope above the platform. The light and the roaring cut off as he
grasped the rope, flashed back into the building, cut off again.
Somewhere somebody had screamed....
Dowland swung about on the rope, went handing himself along it,
back toward the door. His feet flopped about over the floor, unable to
get a stand there for more than an instant. It was a struggle now to
get enough air through the antiradiation field into his lungs. He saw
dust whip past the open door, momentarily obscuring it. The building
bucked with earthquake fury. And where was Trelawney?
He saw the red, wet thing then, lying by the wall just inside the door;
and sickness seized him because Trelawney's body was stretched out
too far to make it seem possible it had ever been that of a man. Dust
blasted in through the door as he reached it, and subsided, leaving a
choking residue trapped within the radiation screen. If he could only
cut off the field....
His gun lay too close to the sodden mess along the wall. Dowland
picked it up, was bending to snatch the climbing harness from the
floor when light flared behind him again. Automatically, he looked
back.
Once more the interior of the building seemed to have split apart.
Wider now. He saw the rushing white current below. To the right,
above the forest on the bank, the sun was a swollen red ball glaring
through layers of mist. And to the left, moving slowly over the river in
the blaze of long-dead daylight, was something both unmistakable
and not to be believed. But, staring at it in the instant before the
scene shivered and vanished again, Dowland suddenly thought he
knew what had happened here.
What he had seen was a spaceship.
He turned, went stumbling hurriedly out the door into the whistling
wind, saw Jill Trelawney standing there, white-faced, eyes huge,
hands to her mouth.
He caught her shoulder. "Come on! We've got to get away from
here."
She gasped, "It—tore him apart!"
"We can't help him...." Dust clouds were spinning over the back of
the mesa, concealing the upper slopes. Dowland glanced to the west,
winced at the towering mountain of darkness sweeping toward them
through the sky. He plunged up the slope, hauling her along behind
him. Jill cried out incoherently once, in a choking voice, but he didn't
stop to hear what she was trying to say. He shoved her into the
house, slammed the door shut behind them, hurried her on down the
hall and into the living room. As they came in, he switched off his AR
field and felt air fill his lungs easily again. It was like surfacing out of
deep water. The detector still hissed its thin warning, but it was
almost inaudible. They would have to risk radiation now.
"Out of your suit, quick! Whatever's happening in the lab has
whistled up a dust storm here. When it hits, that radiation field will
strangle you in a minute outdoors."
She stared at him dumbly.
"Get out of your suit!" Dowland shouted, his nerves snapping. "We're
going down the eastern wall. It's our only chance. But we can't get
down alive if we can't breathe...." Then, as she began unbuckling the
suit hurriedly with shaking fingers, he turned to the pile of camping
equipment beside the fireplace and pawed through it.
He found the communicator and was snapping it to the
mountaineering harness when the front door slammed. He wheeled
about, startled. Jill's radiation suit lay on the floor near the entry hall.
She was gone.
He was tearing the door open three seconds later, shouted, and saw
her through the dust forty feet away, running up toward the forest.
He mightn't have caught her if she hadn't stumbled and gone
headlong. Dowland was on top of her before she could get up. She
fought him in savage silence like an animal, tearing and biting, her
eyes bloodshot slits. There was a mechanical fury about it that
appalled him. But at last he got his right arm free, and brought his
fist up solidly to the side of her jaw. Jill's head flew back, and her
eyes closed.

He came padding up to the eastern side of the mesa with her

minutes later. Here, beyond the ranch area, the ground was bare
rock, with occasional clusters of stunted bushes. The dust had
become blinding, though the main storm was still miles away. There
was no time to stop off at the house to look for the quiz-gun, though
it would have been better to try the descent with a dazed and half-
paralyzed young woman than with the twisting lunatic Jill might turn
into again when she recovered from his punch. At least, he'd have
her tied up. Underfoot were grinding and grumbling noises now, the
ground shaking constantly. At moments he had the feeling of
plodding through something yielding, like quicksand. Only the feeling,
he told himself; the rock was solid enough. But....
Abruptly, he was at the mesa's edge. Dowland slid the girl to the
ground, straightened up, panting, to dab at his smarting eyes. The
mesa behind them had almost vanished in swirling dust.
And through the dust Dowland saw something coming over the open
ground he had just traversed.
He stared at it, mouth open, stunned with a sense of unfairness. The
gigantic shape was still only partly visible, but it was obvious that it
was following them. It approached swiftly over the shaking ground.
Dowland took out his gun, with the oddly calm conviction that it
would be entirely useless against their pursuer. But he brought it up
slowly and leveled it, squinting with streaming eyes through the dust.
And then it happened. The pursuer appeared to falter. It moved again
in some manner; something thundered into the ground beside
Dowland. Then, writhing and twisting—slowly at first, then faster—
the dust-veiled shape seemed to be sinking downward through the
rock surface of the mesa.
In another instant, it was gone.
Seconds passed before Dowland gradually lowered the gun again.
Dazedly, he grew aware of something else that was different now. A
miniature human voice appeared to be jabbering irritably at him from
some point not far away. His eyes dropped to the little communicator
attached to his harness.
The voice came from there.
Terra's grid-power had returned to Lion Mesa.

A week later, Lieutenant Frank Dowland was shown into the office of
the chief of the Solar Police Authority. The chief introduced him to the
two other men there, who were left unidentified, and told him to be
seated.
"Lieutenant," he said, "these gentlemen have a few questions to ask
you. You can speak as openly to them as you would to me."
Dowland nodded. He had recognized one of the gentlemen
immediately—Howard Camhorn, the Coordinator of Research.
Reputedly one of the sharpest minds in the Overgovernment's top
echelons. The other one was unfamiliar. He was a few years younger
than Camhorn, around six inches shorter, chunky, with black hair,
brown eyes, an expression of owlish reflectiveness. Probably,
Dowland thought, wearing contact lenses. "Yes, sir," he said to the
chief, and looked back at the visitors.
"We've seen your report on your recent visit to Terra, Lieutenant
Dowland," Camhorn began pleasantly. "An excellent report,
incidentally—factual, detailed. What we should like to hear now are
the things that you, quite properly, omitted from it. That is, your
personal impressions and conclusions."
"For example," the other man took up, as Dowland hesitated, "Miss
Trelawney has informed us her uncles were attempting to employ the
Ym-400 they had acquired to carry out a time-shift to an earlier Earth
period—to the period known as the Pleistocene, to be somewhat
more exact. From what you saw, would you say they had succeeded
in doing it?"
"I don't know, sir," Dowland said. "I've been shown pictures
representing that period during the past few days. The scene I
described in the report probably might have existed at that time." He
smiled briefly. "However, I have the impression that the very large
flying creature I reported encountering that night is regarded as
being ... well, er ... ah...."
"A product of excited nerves?" the short man said, nodding. "Under
such extraordinary circumstances, that would be quite possible, you
know."
"Yes, sir, I know."
The short man smiled. "But you don't think it was that?"
"No, sir," Dowland said. "I think that I have described exactly what I
did hear and see."
"And you feel the Trelawneys established contact with some previous
Earth period—not necessarily the Pleistocene?"
"Yes, I do."
"And you report having seen a spaceship in that prehistorical
period...."
Dowland shook his head. "No, sir. At the moment I was observing it, I
thought it was that. What I reported was having seen something that
looked like a spaceship."
"What do you think it was?"
"A timeship—if there is such a word."
"There is such a word," Camhorn interrupted lazily. "I'm curious to
hear, lieutenant, what brought you to that conclusion."
"It's a guess, sir. But the thing has to fit together somehow. A
timeship would make it fit."
"In what way?"
"I've been informed," Dowland said, "that the Overgovernment's
scientists have been unable to make a practical use of YM because
something has invariably gone wrong when they did try to use it. I
also heard that there was no way of knowing in advance what would
happen to make an experiment fail. But something always would
happen, and frequently a number of people would get killed."
Camhorn nodded. "That is quite true."
"Well, then," Dowland said, "I think there is a race of beings who
aren't quite in our time and space. They have YM and use it, and
don't want anyone else to use it. They can tell when it's activated
here, and use their own YM to interfere with it. Then another
experiment suddenly turns into a failure.

"But they don't know yet who's using it. When the Trelawneys turned
on their machine, these beings spotted the YM stress pattern back
there in time. They went to that point and reinforced the time-
blending effect with their own YM. The Trelawneys hadn't intended a
complete contact with that first test. The aliens almost succeeded in
blending the two periods completely in the area near the laboratory."
"For what purpose?" Camhorn asked.
"I think they're very anxious to get us located."
"With unfriendly intentions?"
"The ones we ran into didn't behave in a friendly manner. May I ask a
question, sir?"
"Of course," Camhorn said.
"When the Trelawneys' machine was examined, was the supply of YM
adequately shielded?"
"Quite adequately," Camhorn said.
"But when I opened the door, the laboratory was hot. And Miguel
Trelawney died of radiation burns...."
Camhorn nodded. "Those are facts that give your theory some
substance, lieutenant. No question about it. And there is the
additional fact that after you shut off the YM flow in the laboratory,
nearly ten minutes passed before the apparent contact between two
time periods was broken. Your report indicates that the phenomena
you described actually became more pronounced immediately after
the shutoff."
"Yes, sir."
"As if the aliens might have been making every effort to retain
contact with our time?"
"Yes, sir," Dowland said. "That was my impression."
"It's quite plausible. Now, the indications are that Paul Trelawney
actually spent considerable time—perhaps twelve to fourteen hours,
at any rate—in that other period. He gave no hint of what he
experienced during those hours?"
"No, sir, except to say that it was night when he appeared there. He
may have told Miss Trelawney more."
"Apparently, he didn't," Camhorn said. "Before you and he went into
the laboratory, he warned her to watch for the approach of a creature
which answers the description of the gigantic things you encountered
twice. But that was all. Now, here again you've given us your
objective observations. What can you add to them—on a perhaps
more speculative basis?"
"Well, sir," Dowland said, "my opinions on that are, as a matter of
fact, highly speculative. But I think that Paul Trelawney was captured
by the aliens as soon as he appeared in the other time period, and
was able to escape from them a number of hours later. Two of the
aliens who were attempting to recapture him eventually followed him
out on Lion Mesa through another opening the YM stresses had
produced between the time periods, not too far away from the first."
Camhorn's stout companion said thoughtfully, "You believe the
birdlike creature you saw arrived by the same route?"
"Yes, sir," Dowland said, turning to him. "I think that was simply an
accident. It may have been some kind of wild animal that blundered
into the contact area and found itself here without knowing what had
occurred."
"And you feel," the other man went on, "that you yourself were
passing near that contact point in the night at the time you seemed
to be smelling a swamp?"
Dowland nodded. "Yes, sir, I do. Those smells might have been an
illusion, but they seemed to be very distinct. And, of course, there
are no swamps on the mesa itself."

Camhorn said, "We'll assume it was no illusion. It seems to fit into

the general picture. But, lieutenant, on what are you basing your
opinion that Paul Trelawney was a captive of these beings for some
time?"
"There were several things, sir," Dowland said. "One of them is that
when Miss Trelawney regained consciousness in the hospital she
didn't remember having made an attempt to get away from me."
Camhorn nodded. "That was reported."
"She made the attempt," Dowland went on, "immediately after she
had taken off her radiation suit to avoid being choked in the dust
storm on the way down from the mesa. That is one point."
"Go ahead," Camhorn said.
"Another is that when I discovered Paul Trelawney early in the
morning, he was wearing his radiation suit. Judging by his
appearance, he had been in it for hours—and a radiation suit, of
course, is a very inconvenient thing to be in when you're hiking
around in rough country."
"He might," the stout man suggested, "have been afraid of running
into a radioactive area."
Dowland shook his head. "No, sir. He had an instrument which would
have warned him if he was approaching one. It would have made
much more sense to carry the suit, and slip into it again if it became
necessary. I didn't give the matter much thought at the time. But
then the third thing happened. I did not put that in the report
because it was a completely subjective impression. I couldn't prove
now that it actually occurred."
Camhorn leaned forward. "Go ahead."
"It was just before the time periods separated and the creature that
was approaching Miss Trelawney and myself seemed to drop through
the top of the mesa—I suppose it fell back into the other period. I've
described it. It was like a fifty-foot gray slug moving along on its tail
... and there were those two rows of something like short arms. It
wasn't at all an attractive creature. I was frightened to death. But I
was holding a gun—the same gun with which I had stopped another
of those things when it chased me during the night. And the trouble
was that this time I wasn't going to shoot."
"You weren't going to shoot?" Camhorn repeated.
"No, sir. I had every reason to try to blow it to pieces as soon as I
saw it. The other one didn't follow up its attack on me, so it probably
was pretty badly injured. But while I knew that, I was also simply
convinced that it would be useless to pull the trigger. That's as well
as I can explain what happened....
"I think these aliens can control the minds of other beings, but can't
control them through the interference set up by something like our
AR fields. Paul Trelawney appeared in the other time period almost in
their laps. He had a rifle strapped over his back, but presumably they
caught him before he had a chance to use it. They would have
examined him and the equipment he was carrying, and when they
took off his radiation suit, they would have discovered he belonged to
a race which they could control mentally. After that, there would have
been no reason for them to guard him too closely. He was helpless.

"I think Trelawney realized this, and used a moment when his actions
were not being controlled to slip back into the suit. Then he was free
to act again. When they discovered he had escaped, some of them
were detailed to search for him, and two of those pursuers came out
here in our time on the mesa.
"As for Miss Trelawney—well, obviously she wasn't trying to get away
from me. Apparently, she wasn't even aware of what she was doing.
She was simply obeying physically the orders her mind began to
receive as soon as she stepped out of the radiation suit. They would
have been to come to the thing, wherever it was at the moment—
somewhere up to the north of the ranch area, judging from the
direction in which she headed."
There was silence for some seconds. Then Camhorn's companion
observed, "There's one thing that doesn't quite fit in with your theory,
lieutenant."
"What's that, sir?"
"Your report states that you switched off your AR field at the same
time you advised Miss Trelawney to get out of her suit. You should
have been equally subject to the alien's mental instructions."
"Well," Dowland said, "I can attempt to explain that, sir, though again
there is no way to prove what I think. But it might be that these
creatures can control, only one mind at a time. The alien may not
have realized that I had ... well ... knocked Miss Trelawney
unconscious and that she was unable to obey its orders, until it came
to the spot and saw us. My assumption is that it wasn't till that
moment that it switched its mental attack to me."

The stout man—his name was Laillard White, and he was one of
Research's ace trouble-shooters in areas more or less loosely related
to psychology—appeared morosely reflective as he and Camhorn left
Solar Police Authority Headquarters, and turned toward the adjoining
Overgovernment Bureau.
"I gather from your expression," Camhorn remarked, "that our
lieutenant was telling the truth."
White grunted. "Of course, he was—as he saw it."
"And he's sane?"
"Quite sane," White agreed absently.
Camhorn grinned. "Then what's the matter, Lolly? Don't you like the
idea of time-travel?"
"Naturally not. It's an absurdity."
"You're blunt, Lolly. And rash. A number of great minds differ with
you about that."
Laillard White said something rude about great minds in general. He
went on, "Was the machine these Trelawneys built found intact?"
Camhorn nodded. "In perfect condition. I found an opportunity to
look it over when it and the others the Freeholders had concealed on
Terra were brought in."
"And these machines are designed to make it possible to move
through time?"
"No question about that. They function in Riemann space, and are
very soundly constructed. A most creditable piece of work, in fact. It's
only regrettable that the Trelawney brothers were wasted on it. We
might have put their talents to better use. Though as it turned
out...." He shrugged.
White glanced over at him. "What are you talking about?" he asked
suspiciously.
"They didn't accomplish time-travel," Camhorn said, "though in
theory they should have. I know it because we have several
machines based on the same principles. The earliest was built almost
eighty years ago. Two are now designed to utilize the YM thrust. The
Trelawney machine is considerably more advanced in a number of
details than its Overgovernment counterparts, but it still doesn't
make it possible to move in time."
"Why not?"
"I'd like to know," Camhorn said. "The appearance of it is that the
reality we live in takes the same dim view of time-travel that you do.
Time-travel remains a theoretical possibility. But in practice—when,
for example, the YM thrust is applied for that purpose—the thrust is
diverted."
White looked bewildered. "But if Paul Trelawney didn't move through
time, what did he do?"
"What's left?" Camhorn asked. "He moved through space, of course."
"Where?"
Camhorn shrugged. "They penetrated Riemann space," he said,
"after harnessing their machine to roughly nineteen thousand times
the power that was available to us before the Ymir series of elements
dropped into our hands. In theory, Lolly, they might have gone
anywhere in the universe. If we'd had the unreasonable nerve to play
around with multikilograms of YM—knowing what happened when
fractional quantities of a gram were employed—we might have had a
very similar experience."
"I'm still just a little in the dark, you know," Laillard White observed
drily, "as to what the experience consisted of."
"Oh, Lieutenant Dowland's theory wasn't at all far off in that respect.
It's an ironic fact that we have much to thank the Trelawneys for.
There's almost no question at all now what the race of beings they
encountered were responsible for the troubles that have plagued us
in the use of YM. They're not the best of neighbors—neighbors in
Riemann space terms, that is. If they'd known where to look for us,
things might have become rather hot. They had a chance to win the
first round when the Trelawneys lit that sixty-eight kilogram beacon
for them. But they made a few mistakes, and lost us again. It's a
draw so far. Except that we now know about as much about them as
they've ever learned about us. I expect we'll take the second round
handily a few years from now."

White still looked doubtful. "Was it one of their planets the

Trelawneys contacted?"
"Oh, no. At least, it would have been an extremely improbable
coincidence. No, the machine was searching for Terra as Terra is
known to have been in the latter part of the Pleistocene period. The
Trelawneys had provided something like a thousand very specific
factors to direct and confine that search. Time is impenetrable, so the
machine had to find that particular pattern of factors in space, and
did. The aliens—again as Lieutenant Dowland theorized—then moved
through Riemann space to the planet where the YM thrust was
manifesting itself so violently. But once there, they still had no way of
determining where in the universe the thrust had originated—even
though they were, in one sense, within shouting distance of Terra,
and two of them were actually on its surface for a time. It must have
been an extremely frustrating experience all around for our friends."
Laillard White said, "Hm-m," and frowned.
Camhorn laughed. "Let it go, Lolly," he said. "That isn't your field,
after all. Let's turn to what is. What do you make of the fact that
Dowland appears to have been temporarily immune to the mental
commands these creatures can put out?"
"Eh?" White said. His expression turned to one of surprise. "But that's
obvious!"
"Glad to hear it," Camhorn said drily.
"Well, it is. Dowland's attitude showed clearly that he suspected the
truth himself on that point. Naturally, he was somewhat reluctant to
put it into words."
"Naturally. So what did he suspect?"
White shook his head. "It's so simple. The first specimen of humanity
the aliens encountered alive was Paul Trelawney. High genius level,
man! It would take that level to nullify our I.Q. tests in the manner
he and his half-brother did. When those creatures were prowling
around on the mesa, they were looking for that kind of mentality.
Dowland's above average, far from stupid. As you say, you like his
theories. But he's no Trelawney. Unquestionably, the aliens in each
case regarded him as some kind of clever domestic animal. The only
reason he's alive is that they weren't taking him seriously."

"That," Camhorn said thoughtfully, "may have changed a number of

things."
"It may, indeed."
"Do we have anything on hand that would block their specific psi
abilities?"
"Oh, surely. If an AR field can stop them, there's nothing to worry
about in that respect. Our human telepaths wouldn't be seriously
hampered by that degree of interference."
"Very good," Camhorn said. "Do you have any theory about the
partial sensory interpretation of the two areas which both Dowland
and Miss Trelawney reported? The matter of being able to hear the
river on the other planet from time to time."
White nodded. "There are several possible explanations for that. For
one thing...."
"Better save it for lunch, Lolly," Camhorn interrupted, glancing at his
watch. "I see I have two minutes left to make the meeting. Anything
else you feel should be brought up at the moment?"
"Just one thing," White said. "If the Trelawneys' machine is capable
of locating a Terra-type planet anywhere in the universe...."
Camhorn nodded. "It is."
"Then," White said, "we've solved our exploding population problem,
haven't we?"
"For the time being, we have," Camhorn agreed. "As a matter of fact,
Lolly, that's precisely what the meeting I'm headed for is about."
"Then the Terran Freeholders can stop worrying about the political
pressures that have threatened to turn Terra into another hygienically
overcrowded slum-world."
"True enough," Camhorn said. "In another few years, if things go
right, every man, woman and child can become a Freeholder—
somewhere."
"So the Trelawneys got what they wanted, after all...."
"They did, in a way. If the brothers knew the whole score, I think
they'd be satisfied. The situation has been explained to their niece.
She is."
THE END
 *** END OF THE PROJECT GUTENBERG EBOOK THE BEACON TO
ELSEWHERE ***

Updated editions will replace the previous one—the old editions will
be renamed.

Creating the works from print editions not protected by U.S.

copyright law means that no one owns a United States copyright in
these works, so the Foundation (and you!) can copy and distribute it
in the United States without permission and without paying
copyright royalties. Special rules, set forth in the General Terms of
Use part of this license, apply to copying and distributing Project
Gutenberg™ electronic works to protect the PROJECT GUTENBERG™
concept and trademark. Project Gutenberg is a registered trademark,
and may not be used if you charge for an eBook, except by following
the terms of the trademark license, including paying royalties for use
of the Project Gutenberg trademark. If you do not charge anything
for copies of this eBook, complying with the trademark license is
very easy. You may use this eBook for nearly any purpose such as
creation of derivative works, reports, performances and research.
Project Gutenberg eBooks may be modified and printed and given
away—you may do practically ANYTHING in the United States with
eBooks not protected by U.S. copyright law. Redistribution is subject
to the trademark license, especially commercial redistribution.

START: FULL LICENSE

Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

textbookfull.com