Novel Sparse-Coded Ambient Backscatter Communication for Massive IoT Connectivity
Extracted source text ยท arXiv 1806.02975
%% bare_jrnl.tex
%% V1.4b
%% 2015/08/26
%% by Michael Shell
%% see http://www.michaelshell.org/
%% for current contact information.
%%
%% This is a skeleton file demonstrating the use of IEEEtran.cls
%% (requires IEEEtran.cls version 1.8b or later) with an IEEE
%% journal paper.
%%
%% Support sites:
%% http://www.michaelshell.org/tex/ieeetran/
%% http://www.ctan.org/pkg/ieeetran
%% and
%% http://www.ieee.org/
%%*************************************************************************
%% Legal Notice:
%% This code is offered as-is without any warranty either expressed or
%% implied; without even the implied warranty of MERCHANTABILITY or
%% FITNESS FOR A PARTICULAR PURPOSE!
%% User assumes all risk.
%% In no event shall the IEEE or any contributor to this code be liable for
%% any damages or losses, including, but not limited to, incidental,
%% consequential, or any other damages, resulting from the use or misuse
%% of any information contained here.
%%
%% All comments are the opinions of their respective authors and are not
%% necessarily endorsed by the IEEE.
%%
%% This work is distributed under the LaTeX Project Public License (LPPL)
%% ( http://www.latex-project.org/ ) version 1.3, and may be freely used,
%% distributed and modified. A copy of the LPPL, version 1.3, is included
%% in the base LaTeX documentation of all distributions of LaTeX released
%% 2003/12/01 or later.
%% Retain all contribution notices and credits.
%% ** Modified files should be clearly indicated as such, including **
%% ** renaming them and changing author support contact information. **
%%*************************************************************************
% *** Authors should verify (and, if needed, correct) their LaTeX system ***
% *** with the testflow diagnostic prior to trusting their LaTeX platform ***
% *** with production work. The IEEE's font choices and paper sizes can ***
% *** trigger bugs that do not appear when using other class files. *** ***
% The testflow support page is at:
% http://www.michaelshell.org/tex/testflow/
\documentclass[journal,comsoc]{IEEEtran}
%
% If IEEEtran.cls has not been installed into the LaTeX system files,
% manually specify the path to it like:
% \documentclass[journal]{../sty/IEEEtran}
% Some very useful LaTeX packages include:
% (uncomment the ones you want to load)
% *** MISC UTILITY PACKAGES ***
%
%\usepackage{ifpdf}
% Heiko Oberdiek's ifpdf.sty is very useful if you need conditional
% compilation based on whether the output is pdf or dvi.
% usage:
% \ifpdf
% % pdf code
% \else
% % dvi code
% \fi
% The latest version of ifpdf.sty can be obtained from:
% http://www.ctan.org/pkg/ifpdf
% Also, note that IEEEtran.cls V1.7 and later provides a builtin
% \ifCLASSINFOpdf conditional that works the same way.
% When switching from latex to pdflatex and vice-versa, the compiler may
% have to be run twice to clear warning/error messages.
% *** CITATION PACKAGES ***
%
%\usepackage{cite}
% cite.sty was written by Donald Arseneau
% V1.6 and later of IEEEtran pre-defines the format of the cite.sty package
% \cite{} output to follow that of the IEEE. Loading the cite package will
% result in citation numbers being automatically sorted and properly
% "compressed/ranged". e.g., [1], [9], [2], [7], [5], [6] without using
% cite.sty will become [1], [2], [5]--[7], [9] using cite.sty. cite.sty's
% \cite will automatically add leading space, if needed. Use cite.sty's
% noadjust option (cite.sty V3.8 and later) if you want to turn this off
% such as if a citation ever needs to be enclosed in parenthesis.
% cite.sty is already installed on most LaTeX systems. Be sure and use
% version 5.0 (2009-03-20) and later if using hyperref.sty.
% The latest version can be obtained at:
% http://www.ctan.org/pkg/cite
% The documentation is contained in the cite.sty file itself.
% *** GRAPHICS RELATED PACKAGES ***
%
\ifCLASSINFOpdf
% \usepackage[pdftex]{graphicx}
% declare the path(s) where your graphic files are
% \graphicspath{{../pdf/}{../jpeg/}}
% and their extensions so you won't have to specify these with
% every instance of \includegraphics
% \DeclareGraphicsExtensions{.pdf,.jpeg,.png}
\else
% or other class option (dvipsone, dvipdf, if not using dvips). graphicx
% will default to the driver specified in the system graphics.cfg if no
% driver is specified.
% \usepackage[dvips]{graphicx}
% declare the path(s) where your graphic files are
% \graphicspath{{../eps/}}
% and their extensions so you won't have to specify these with
% every instance of \includegraphics
% \DeclareGraphicsExtensions{.eps}
\fi
% graphicx was written by David Carlisle and Sebastian Rahtz. It is
% required if you want graphics, photos, etc. graphicx.sty is already
% installed on most LaTeX systems. The latest version and documentation
% can be obtained at:
% http://www.ctan.org/pkg/graphicx
% Another good source of documentation is "Using Imported Graphics in
% LaTeX2e" by Keith Reckdahl which can be found at:
% http://www.ctan.org/pkg/epslatex
%
% latex, and pdflatex in dvi mode, support graphics in encapsulated
% postscript (.eps) format. pdflatex in pdf mode supports graphics
% in .pdf, .jpeg, .png and .mps (metapost) formats. Users should ensure
% that all non-photo figures use a vector format (.eps, .pdf, .mps) and
% not a bitmapped formats (.jpeg, .png). The IEEE frowns on bitmapped formats
% which can result in "jaggedy"/blurry rendering of lines and letters as
% well as large increases in file sizes.
%
% You can find documentation about the pdfTeX application at:
% http://www.tug.org/applications/pdftex
% *** MATH PACKAGES ***
%
%\usepackage{amsmath}
% A popular package from the American Mathematical Society that provides
% many useful and powerful commands for dealing with mathematics.
%
% Note that the amsmath package sets \interdisplaylinepenalty to 10000
% thus preventing page breaks from occurring within multiline equations. Use:
%\interdisplaylinepenalty=2500
% after loading amsmath to restore such page breaks as IEEEtran.cls normally
% does. amsmath.sty is already installed on most LaTeX systems. The latest
% version and documentation can be obtained at:
% http://www.ctan.org/pkg/amsmath
\usepackage[T1]{fontenc}
\usepackage{cite}
\usepackage{graphicx}
\usepackage{epstopdf}
\usepackage{amsmath}
\interdisplaylinepenalty=2500
\usepackage{algorithm}
\usepackage{algpseudocode}
\usepackage{tabularx}
\usepackage{dblfloatfix}
\usepackage{caption}
\usepackage{subcaption}
\usepackage{multirow}
% *** SPECIALIZED LIST PACKAGES ***
%
%\usepackage{algorithmic}
% algorithmic.sty was written by Peter Williams and Rogerio Brito.
% This package provides an algorithmic environment fo describing algorithms.
% You can use the algorithmic environment in-text or within a figure
% environment to provide for a floating algorithm. Do NOT use the algorithm
% floating environment provided by algorithm.sty (by the same authors) or
% algorithm2e.sty (by Christophe Fiorio) as the IEEE does not use dedicated
% algorithm float types and packages that provide these will not provide
% correct IEEE style captions. The latest version and documentation of
% algorithmic.sty can be obtained at:
% http://www.ctan.org/pkg/algorithms
% Also of interest may be the (relatively newer and more customizable)
% algorithmicx.sty package by Szasz Janos:
% http://www.ctan.org/pkg/algorithmicx
% *** ALIGNMENT PACKAGES ***
%
%\usepackage{array}
% Frank Mittelbach's and David Carlisle's array.sty patches and improves
% the standard LaTeX2e array and tabular environments to provide better
% appearance and additional user controls. As the default LaTeX2e table
% generation code is lacking to the point of almost being broken with
% respect to the quality of the end results, all users are strongly
% advised to use an enhanced (at the very least that provided by array.sty)
% set of table tools. array.sty is already installed on most systems. The
% latest version and documentation can be obtained at:
% http://www.ctan.org/pkg/array
% IEEEtran contains the IEEEeqnarray family of commands that can be used to
% generate multiline equations as well as matrices, tables, etc., of high
% quality.
% *** SUBFIGURE PACKAGES ***
%\ifCLASSOPTIONcompsoc
% \usepackage[caption=false,font=normalsize,labelfont=sf,textfont=sf]{subfig}
%\else
% \usepackage[caption=false,font=footnotesize]{subfig}
%\fi
% subfig.sty, written by Steven Douglas Cochran, is the modern replacement
% for subfigure.sty, the latter of which is no longer maintained and is
% incompatible with some LaTeX packages including fixltx2e. However,
% subfig.sty requires and automatically loads Axel Sommerfeldt's caption.sty
% which will override IEEEtran.cls' handling of captions and this will result
% in non-IEEE style figure/table captions. To prevent this problem, be sure
% and invoke subfig.sty's "caption=false" package option (available since
% subfig.sty version 1.3, 2005/06/28) as this is will preserve IEEEtran.cls
% handling of captions.
% Note that the Computer Society format requires a larger sans serif font
% than the serif footnote size font used in traditional IEEE formatting
% and thus the need to invoke different subfig.sty package options depending
% on whether compsoc mode has been enabled.
%
% The latest version and documentation of subfig.sty can be obtained at:
% http://www.ctan.org/pkg/subfig
% *** FLOAT PACKAGES ***
%
%\usepackage{fixltx2e}
% fixltx2e, the successor to the earlier fix2col.sty, was written by
% Frank Mittelbach and David Carlisle. This package corrects a few problems
% in the LaTeX2e kernel, the most notable of which is that in current
% LaTeX2e releases, the ordering of single and double column floats is not
% guaranteed to be preserved. Thus, an unpatched LaTeX2e can allow a
% single column figure to be placed prior to an earlier double column
% figure.
% Be aware that LaTeX2e kernels dated 2015 and later have fixltx2e.sty's
% corrections already built into the system in which case a warning will
% be issued if an attempt is made to load fixltx2e.sty as it is no longer
% needed.
% The latest version and documentation can be found at:
% http://www.ctan.org/pkg/fixltx2e
%\usepackage{stfloats}
% stfloats.sty was written by Sigitas Tolusis. This package gives LaTeX2e
% the ability to do double column floats at the bottom of the page as well
% as the top. (e.g., "\begin{figure*}[!b]" is not normally possible in
% LaTeX2e). It also provides a command:
%\fnbelowfloat
% to enable the placement of footnotes below bottom floats (the standard
% LaTeX2e kernel puts them above bottom floats). This is an invasive package
% which rewrites many portions of the LaTeX2e float routines. It may not work
% with other packages that modify the LaTeX2e float routines. The latest
% version and documentation can be obtained at:
% http://www.ctan.org/pkg/stfloats
% Do not use the stfloats baselinefloat ability as the IEEE does not allow
% \baselineskip to stretch. Authors submitting work to the IEEE should note
% that the IEEE rarely uses double column equations and that authors should try
% to avoid such use. Do not be tempted to use the cuted.sty or midfloat.sty
% packages (also by Sigitas Tolusis) as the IEEE does not format its papers in
% such ways.
% Do not attempt to use stfloats with fixltx2e as they are incompatible.
% Instead, use Morten Hogholm'a dblfloatfix which combines the features
% of both fixltx2e and stfloats:
%
% \usepackage{dblfloatfix}
% The latest version can be found at:
% http://www.ctan.org/pkg/dblfloatfix
%\ifCLASSOPTIONcaptionsoff
% \usepackage[nomarkers]{endfloat}
% \let\MYoriglatexcaption\caption
% \renewcommand{\caption}[2][\relax]{\MYoriglatexcaption[#2]{#2}}
%\fi
% endfloat.sty was written by James Darrell McCauley, Jeff Goldberg and
% Axel Sommerfeldt. This package may be useful when used in conjunction with
% IEEEtran.cls' captionsoff option. Some IEEE journals/societies require that
% submissions have lists of figures/tables at the end of the paper and that
% figures/tables without any captions are placed on a page by themselves at
% the end of the document. If needed, the draftcls IEEEtran class option or
% \CLASSINPUTbaselinestretch interface can be used to increase the line
% spacing as well. Be sure and use the nomarkers option of endfloat to
% prevent endfloat from "marking" where the figures would have been placed
% in the text. The two hack lines of code above are a slight modification of
% that suggested by in the endfloat docs (section 8.4.1) to ensure that
% the full captions always appear in the list of figures/tables - even if
% the user used the short optional argument of \caption[]{}.
% IEEE papers do not typically make use of \caption[]'s optional argument,
% so this should not be an issue. A similar trick can be used to disable
% captions of packages such as subfig.sty that lack options to turn off
% the subcaptions:
% For subfig.sty:
% \let\MYorigsubfloat\subfloat
% \renewcommand{\subfloat}[2][\relax]{\MYorigsubfloat[]{#2}}
% However, the above trick will not work if both optional arguments of
% the \subfloat command are used. Furthermore, there needs to be a
% description of each subfigure *somewhere* and endfloat does not add
% subfigure captions to its list of figures. Thus, the best approach is to
% avoid the use of subfigure captions (many IEEE journals avoid them anyway)
% and instead reference/explain all the subfigures within the main caption.
% The latest version of endfloat.sty and its documentation can obtained at:
% http://www.ctan.org/pkg/endfloat
%
% The IEEEtran \ifCLASSOPTIONcaptionsoff conditional can also be used
% later in the document, say, to conditionally put the References on a
% page by themselves.
% *** PDF, URL AND HYPERLINK PACKAGES ***
%
%\usepackage{url}
% url.sty was written by Donald Arseneau. It provides better support for
% handling and breaking URLs. url.sty is already installed on most LaTeX
% systems. The latest version and documentation can be obtained at:
% http://www.ctan.org/pkg/url
% Basically, \url{my_url_here}.
% *** Do not adjust lengths that control margins, column widths, etc. ***
% *** Do not use packages that alter fonts (such as pslatex). ***
% There should be no need to do such things with IEEEtran.cls V1.6 and later.
% (Unless specifically asked to do so by the journal or conference you plan
% to submit to, of course. )
% correct bad hyphenation here
\hyphenation{op-tical net-works semi-conduc-tor}
\begin{document}
%
% paper title
% Titles are generally capitalized except for words such as a, an, and, as,
% at, but, by, for, in, nor, of, on, or, the, to and up, which are usually
% not capitalized unless they are the first or last word of the title.
% Linebreaks \\ can be used within to get better formatting as desired.
% Do not put math or special symbols in the title.
\title{Novel Sparse-Coded Ambient Backscatter \\Communication for Massive IoT Connectivity}
%
%
% author names and IEEE memberships
% note positions of commas and nonbreaking spaces ( ~ ) LaTeX will not break
% a structure at a ~ so this keeps an author's name from being broken across
% two lines.
% use \thanks{} to gain access to the first footnote area
% a separate \thanks must be used for each paragraph as LaTeX2e's \thanks
% was not built to handle multiple paragraphs
%
\author{Tae Yeong Kim and Dong In Kim% <-this % stops a space
\thanks{T. Y. Kim and D. I. Kim are with the School
of Information and Communication Engineering, Sungkyunkwan University (SKKU), Suwon 16419, Korea (e-mail: \{tangoeye,$\,$dongin\}@skku.edu).}}
% note the % following the last \IEEEmembership and also \thanks -
% these prevent an unwanted space from occurring between the last author name
% and the end of the author line. i.e., if you had this:
%
% \author{....lastname \thanks{...} \thanks{...} }
% ^------------^------------^----Do not want these spaces!
%
% a space would be appended to the last name and could cause every name on that
% line to be shifted left slightly. This is one of those "LaTeX things". For
% instance, "\textbf{A} \textbf{B}" will typeset as "A B" not "AB". To get
% "AB" then you have to do: "\textbf{A}\textbf{B}"
% \thanks is no different in this regard, so shield the last } of each \thanks
% that ends a line with a % and do not let a space in before the next \thanks.
% Spaces after \IEEEmembership other than the last one are OK (and needed) as
% you are supposed to have spaces between the names. For what it is worth,
% this is a minor point as most people would not even notice if the said evil
% space somehow managed to creep in.
% The paper headers
%\markboth{Journal of \LaTeX\ Class Files,~Vol.~14, No.~8, August~2015}%
%{Shell \MakeLowercase{\textit{et al.}}: Bare Demo of IEEEtran.cls for IEEE Journals}
% The only time the second header will appear is for the odd numbered pages
% after the title page when using the twoside option.
%
% *** Note that you probably will NOT want to include the author's ***
% *** name in the headers of peer review papers. ***
% You can use \ifCLASSOPTIONpeerreview for conditional compilation here if
% you desire.
% If you want to put a publisher's ID mark on the page you can do it like
% this:
%\IEEEpubid{0000--0000/00\$00.00~\copyright~2015 IEEE}
% Remember, if you use this you must call \IEEEpubidadjcol in the second
% column for its text to clear the IEEEpubid mark.
% use for special paper notices
%\IEEEspecialpapernotice{(Invited Paper)}
% make the title area
\maketitle
% As a general rule, do not put math, special symbols or citations
% in the abstract or keywords.
\begin{abstract}
Low-power ambient backscatter communication (AmBC) relying on radio-frequency (RF) energy harvesting is an energy-efficient solution for batteryless Internet of things (IoT). However, ambient backscatter signals are severely faded by dyadic backscatter channel (DBC), limiting connectivity in conventional orthogonal time-division-based AmBC (TD-AmBC).
In order to support massive connectivity in AmBC, we propose sparse-coded AmBC (SC-AmBC) based on non-orthogonal signaling.
Sparse code utilizes inherent sparsity of AmBC where power supplies of RF tags rely on ambient RF energy harvesting.
Consequently, sparse-coded backscatter modulation algorithm (SC-BMA) can enable non-orthogonal multiple access (NOMA) as well as $M$-ary modulation for concurrent backscatter transmissions, providing additional diversity gain.
These sparse codewords from multiple tags can be efficiently detected at access point (AP) using iterative message passing algorithm (MPA).
To overcome DBC along with intersymbol interference (ISI), we propose dyadic channel estimation algorithm (D-CEA) and dyadic MPA (D-MPA) exploiting weighted-sum of the ISI for information exchange in factor graph.
Simulation results validate the potential of the SC-AmBC in terms of connectivity, detection performance and sum throughput.
\end{abstract}
% Note that keywords are not normally used for peerreview papers.
\begin{IEEEkeywords}
RF energy harvesting, ambient backscatter communication (AmBC), sparse code multiple access (SCMA), dyadic backscatter channel (DBC), iterative decoder.
\end{IEEEkeywords}
% For peer review papers, you can put extra information on the cover
% page as needed:
% \ifCLASSOPTIONpeerreview
% \begin{center} \bfseries EDICS Category: 3-BBND \end{center}
% \fi
%
% For peerreview papers, this IEEEtran command inserts a page break and
% creates the second title. It will be ignored for other modes.
\IEEEpeerreviewmaketitle
\section{Introduction}
% The very first letter is a 2 line initial drop letter followed
% by the rest of the first word in caps.
%
% form to use if the first word consists of a single letter:
% \IEEEPARstart{A}{demo} file is ....
%
% form to use if you need the single drop letter followed by
% normal text (unknown if ever used by the IEEE):
% \IEEEPARstart{A}{}demo file is ....
%
% Some journals put the first two words in caps:
% \IEEEPARstart{T}{his demo} file is ....
%
% Here we have the typical use of a "T" for an initial drop letter
% and "HIS" in caps to complete the first word.
% Background
% Massive connectivity via AmBC
\IEEEPARstart{M}{assive} connectivity is a key component of future Internet of things (IoT) where radio frequency (RF) devices continuously interact with humans' daily routine.
The vision includes smart home gadgets connected to Wi-Fi (Wireless Fidelity) \cite{passive,backfi}, real-time interactive applications such as Bluetooth exchanging video/audio data with smart phone \cite{ble}, and inventory management in huge warehouse \cite{drone}.
In these applications, RF devices should be designed in low power for prolonging their lifetime and small-form factor for saving cost.
Unfortunately, it is reported that most of the power in RF devices are consumed by oscillator which generates sinusoidal carrier signals for modulation of data. The high power consumption for this digital-to-analog converter (DAC) hinders connectivity for numerous IoT applications with energy-constrained RF devices \cite{ambB,wifiB,turbo}.
Hence, alternative modulation without the oscillator is considered to be a promising solution for fulfilling IoT requirements.
% Small RF Hardware
Ambient backscatter communication (AmBC) is an innovative solution for batteryless IoT, since it replaces power-hungry oscillator with passive load impedances connected to RF antenna \cite{ambB,wifiB,turbo}.
RF tags, sort of RF devices composed of low-power micro-controller and load impedances, can reuse ambient signals such as television (TV) \cite{ambB,turbo} and Wi-Fi \cite{passive,backfi,wifiB} for sending their data.
Instead of generating RF sinusoidal signals, RF tags simply reflect incoming ambient signals and modulate their data by switching load impedances.
These load impedances implemented by passive circuit components, resistors and capacitors can constitute 2-dimensional signal constellation space spanned by real and imaginary axes of reflection coefficients \cite{16qam,qam}. Only a few $\mu\mathrm{W}$ of power is required for AmBC, which is in sharp contrast to traditional active RF radios demanding few tens of $\mathrm{mW}$.
In addition, ultra-low power consumption of backscatter tags can sustain with ambient RF energy harvesting (EH). Ambient RF signals can be recycled not only for EH but also for information transmission over RF signals by controlling reflection coefficients at RF tags. Therefore, AmBC has strong advantages in terms of energy efficiency and installation cost by reducing size of RF hardware.
\subsection{Related Works}
Despite of its benefits, AmBC has several challenges which are originated from inherent characteristic of RF signals and channels.
% Signal
Firstly, the ambient signals are not dedicated to backscatter communications. Hence, some of backscatter schemes based on unmodulated excitation signals such as energy beamforming \cite{retro,RelayB}, space-time coding \cite{dyadic5, dyadic6}, frequency up/down conversion in radio-frequency identification (RFID) \cite{ble, drone} may not be available in AmBC.
Without dedicated RF infrastructure (i.e., 915 MHz RFID reader and power beacon), previous works such as LoRa \cite{LoRaB}, rateless code \cite{rateless}, statistical inference \cite{inference}, time-hopping spreading spectrum \cite{TH} are not directly applied to AmBC scenarios.
Since low-rate symbols are modulated on high-rate ambient signals, modulations and detections should be redesigned for applications of AmBC.
% Channel
Second, channel propagation of AmBC is distinguished from that of active radios \cite{dyadic1,dyadic2,dyadic3,dyadic4,dyadic5,dyadic6}.
Compared to one-way channel between transmitter and receiver \cite{wireless}, backscatter radios experience two-way channel \footnote{Here we assume {\em monostatic} ambient backscatter where the reader (e.g., Wi-Fi AP) generates RF signals intended for legacy (Wi-Fi) users (instead of RF tags) and also acts as the receiver for the signals reflected from RF tags \cite{backfi,ofdm2}. On the other hand, there is {\em bistatic} ambient backscatter where the RF signals are generated from the separated TV or FM station, and cellular base station, while the reader receiving the reflected signals from RF tags \cite{ambB, turbo}.}: forward channel from reader to tag; backward channel from tag to reader. The cascade of two channels is referred to as {\em dyadic backscatter channel (DBC)} and exhibits significant channel attenuations \cite{dyadic1} as well as delay spreads \cite{ofdm2} compared to the one-way channel. Eventually, the phenomena leads to low signal-to-noise ratio (SNR) for detection and severe intersymbol interference (ISI).
To overcome the inherent limitations in AmBC, various modulation and detection techniques based on diversity combining schemes are proposed.
% Multi-antenna
Multiple receive antennas can enhance detection performance in AmBC. For example, interference cancellation schemes using channel differences \cite{turbo}, or zero-forcing \cite{cooperative} are proposed.
The multi-antenna configurations are useful for achieving spatial diversity gain and effective as the number of receive antennas increases.
% Spreading Sequences
On the other hand, modulation and detection based on spreading sequences can also improve reliability achieving temporal diversity gain.
In \cite{backfi, ambB}, detection SNR of AmBC is reinforced using maximal-ratio combining of backscatter samples.
In \cite{ofdm1}, periodicity of ambient signal is utilized using cyclic prefix (CP) in orthogonal frequency division multiplexing (OFDM) sequences. Orthogonal spreading structure similar to code-division multiple access (CDMA) is proposed in \cite{turbo}.
% Sparsity+Spreading Sequence
However, most of AmBC assumed orthogonality in both modulations and detections. For instance, non-overlapping backscatter symbols are typically represented on 2-dimensional Euclidean space (e.g., on-off keying (OOK) \cite{ambB, wifiB}, phase-shift keying (PSK) \cite{backfi}, ternary encoding \cite{codingB}) with orthogonal multiple access (OMA) such as time-division multiple access (TDMA) \cite{ambB-TD} for providing connectivity. For detections, conventional AmBC ignored ISI in DBC by assuming long period of backscatter symbol \cite{ofdm2} or simply filtered it, which reduces SNR for signal detections \cite{backfi, ofdm2}.
In AmBC, tag backscatters data in active mode and harvests RF energy in idle mode whose time portion is much larger than the active mode \cite{rateless}. Consequently, we can utilize the duty-cycling operation by employing {\em sparse coding} over ambient RF signals and support non-orthogonal multiple access (NOMA) for massive connectivity. In addition, low-rate backscatter data are delivered over high-rate ambient RF signal, which can be equivalently regarded as {\em spreading sequences} to provide diversity gain for reliable backscatter communications \cite{ofdm1}. Based on these observations, we can design non-orthogonal sparse coding and detector to realize massive connectivity for AmBC.
\subsection{Proposed Sparse-Coded AmBC (SC-AmBC)}
On the other hand, non-orthogonal signaling in both modulation and detection is proposed recently.
% SCMA
For example, sparse code multiple access (SCMA) proposed in Huawei company proved effectiveness in highly-overloaded network \cite{scma1,scma2}. Especially, sparse coding has three desirable properties which are suitable for massive connectivity:
\begin{itemize}
\item {\bf Grant-Free Uplink Transmission:} With sparse coding, RF tag can transmit data in the uplink based on predefined sparse codebook. The codebook structure can accommodate massive and bursty uplink traffic without request-grant procedure \cite{grant-free}. The property can lead to low-latency without scheduling and resource allocation in contrast to the cognitive radio based AmBC \cite{ambB-TD}.
\item {\bf NOMA:} Sparse code encodes data in multi-dimensional complex codewords providing higher degree of freedoms than conventional 2-dimensional AmBC modulators \cite{ambB, wifiB, backfi, codingB}. As a result, sparse code can support concurrent RF transmissions and multiple-access interference (MAI) can be efficiently resolved by iterative message passing algorithm (MPA) \cite{scma1}.
\item {\bf Projected $M$-ary Modulation:} In conventional AmBC modulators, $M$-ary modulation is only implemented by $M$ distinct load impedances increasing form factors of RF devices for higher-order modulations \cite{backfi, 16qam, qam}. On the other hand, with sparse coding, $M$-ary data can be represented by few constellation symbols using codeword projection/expansion method \cite{iter-Mary,code-project}. Thus, $M$-ary modulation can be realized in AmBC without increasing form factors of RF devices.
\end{itemize}
In addition, due to increasing network densification and delay spreads in multipath propagation, exploiting ISI can achieve better spectral properties than conventional detectors by removing CP \cite{FBMC}. Based on these observations, we propose sparse-coded AmBC (SC-AmBC) including design of modulator and detector for DBC.
Our main contributions can be summarized as follows:
\begin{itemize}
\item Sparse-coded backscatter modulation algorithm (SC-BMA) supporting $M$-ary modulation for AmBC with small-form factors.
\item Dyadic channel estimation algorithm (D-CEA) fully capturing characteristics of DBC.
\item Dyadic MPA (D-MPA) based on dyadic factor graph exploiting the weighted-sum of ISI caused by DBC in information exchange process.
\end{itemize}
To the best of the authors' knowledge, implementing the sparse code for energy-constrained AmBC is not fully explored in literature.
The rest of the paper is organized as follows. Section \ref{sec:system} describes SC-AmBC and motivating examples of our model therein. Section \ref{sec:enc} presents the load modulation including sparse coding algorithm and channel estimation algorithm, and iterative detector based on DBC is proposed in Section \ref{sec:dec} . Simulation results are provided in Section \ref{sec:simulation}. Finally, Section \ref{sec:conclusion} concludes the paper.
\begin{figure*}
\centering
\includegraphics[width=13cm]{System}
\caption{Proposed AmBC network with single AP and multiple RF tags employing ambient RF energy harvesting and backscatter communications using sparse code.}
\label{fig:system}
\end{figure*}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{System Model}\label{sec:system}
\subsection{Notation}
% vector & matrix related
Matrices (vectors) are denoted with upper (lower) case bold-face letters (e.g. $\mathbf{A}$ or $\mathbf{a}$); the operator $\left ( \cdot \right )^{H}$ denotes the Hermitian; the operator $\left \| \cdot \right \|$ denotes the vector norm; the operator $\otimes$ denotes the Kronecker product; the set of real, complex and binary numbers are respectively denoted by $\mathbb{R}$, $\mathbb{C}$ and $\mathbb{B}$; the vector $\mathbf{1}_k$ denotes all-one column vector with size $k$ and the vector $\mathbf{0}_k$ denotes all-zero column vector with size $k$;
% probability related
the operator $p \left ( \mathcal{A} \right )$ denotes the probability of the event $\mathcal{A}$. the operator $\mathbb{E}\left ( \cdot \right )$ denotes statistical expectation; $\mathcal{CN}(\mu,\sigma^2)$ denotes the circularly symmetric complex normal distribution with mean $\mu$ and variance $\sigma^2$.
\subsection{System Model}
% Wi-Fi AP
We consider the AmBC network \cite{ABCN-kaibin} based on access point (AP) such as Wi-Fi as depicted in Fig.~\ref{fig:system}.
The AP transmits data to its client as well as energy to nearby RF tags \cite{16qam,qam,codedqam, backfi}, replacing expensive and high-power RFID reader.
In terms of deployment cost, utilizing ambient signals is very attractive for backscatter communication as they exist everywhere for providing low-cost Internet connectivity (e.g., TV signal, Wi-Fi, BlueTooth).
% RF Tag
In AP's coverage area, there are multiple RF tags utilizing the ambient signals transmitted from the AP to downlink (DL) client.
These RF tags can use the ambient signals for RF energy harvesting using rectifier circuit exempting large-sized battery.
If they harvest sufficient energy for activating backscatter communication, they modulate load impedances connected to micro-controller to transmit $M$-ary data instead of large and power-intensive oscillator component \cite{16qam,qam,codedqam, backfi}.
Depending on ambient energy harvesting, tags are only activated in small fraction of time and mostly remain idle to harvest sufficient energy.
The sparsity of ambient backscatter signals motivates us encoding data to multi-dimensional complex codewords \cite{scma1, scma2, scma3, dyadic3} selected by time-variant reflection coefficients \cite{hybridB, dyadic3} and enables low-complexity decoding even for massive number of concurrent RF transmissions and higher-order modulation.
Consequently, SC-BMA among RF tags can enable NOMA, while providing massive connectivity for low-cost RF tags.
% Detection
Upon the tags' modulation, AP receives low-rate uplink (UL) backscatter signals from the tags while transmitting high-rate DL signals to the client. Assuming single antenna at AP where transmission and reception occur simultaneously in full-duplex, these signals are overlapped causing self-interference (SI) \cite{backfi, SIC}.
In addition, high-sampling rate of DL signals significantly reduces sampling period of UL backscatter signals, causing ISI in multipath propagation environments \cite{ofdm2}.
Moreover, non-orthogonality of sparse code may cause MAI among active tags, degrading detection performance of concurrent RF transmissions \cite{scma1, scma2, scma3, scma4, log-MAP, LDS}.
Therefore, the detection of backscatter signals requires mitigation of the three interferences (i.e., SI, ISI and MAI). Based on self-interference cancellation (SIC) \cite{SIC}, strong DL interference can be efficiently removed so that UL backscatter detection can be feasible in practice \cite{backfi}.
Then, D-CEA followed by low-complexity iterative D-MPA is newly implemented to effectively detect data in the presence of the ISI and MAI.
Clearly, the ambient backscatter rather exploiting interferences significantly improves spectral efficiency in three-folds: full-duplex \cite{backfi, SIC}; NOMA with sparse code \cite{scma1, scma2, scma3}; multipath-diversity combining \cite{wireless}.
\begin{figure}
\centering
\includegraphics[width=8.9 cm]{Protocol}
\caption{Link layer protocol for AmBC with sparse code.}
\label{fig:protocol}
\end{figure}
To support multiple RF tags, link layer protocol of AmBC is composed of channel estimation and data transmission phase as described in Fig.~\ref{fig:protocol}. In channel estimation phase, AP estimates the SI channel using silent period for RF tags \cite{backfi}. Then, the tags transmit sparse pseudo random symbols known to AP, and it can efficiently estimate backscatter channels using compressed sensing algorithms \cite{rateless, CS}.
Moreover, as the DBC \cite{dyadic1,dyadic2,dyadic3,dyadic4,dyadic5,dyadic6} experiences two-way fading in contrast to the conventional one-way fading models, D-CEA capturing the characteristics of these channels will be described in Section \ref{sec:enc}.
In data transmission phase, ambient signal from AP can be sliced into $K$ time slots, each with $L$ samples, to spread $N \geq K$ backscatter signals to increase signal-to-noise ratio (SNR) and reduce circuit-power consumption \cite{backfi, ambB, wifiB, turbo, ambB-analysis, ambB-noncoherent, codingB}. As RF tags employ predefined codebooks for data transmissions, active tags can immediately transmit data with low latency without resource allocation and scheduling procedures \cite{grant-free}.
At time slot $k\in \{1,2,\cdots,K\}$, the ambient signal with average transmit power $\sigma_s^2$ can be expressed by
\begin{eqnarray}
{\mathbf{s}}_k=[s_{1,k},\cdots,s_{L,k}]^T\in \mathbb{C}^{L\times 1}, \quad
s_{l,k}\sim \mathcal{CN}(0,\sigma_s^2),
\end{eqnarray}
where Gaussian source model with sampling period $T_s$ is assumed to design ambient signal \cite{ambB-noncoherent}.
% Forward Channel Model
Next, the signal broadcasted from the AP is attenuated by forward channels. The discrete-time channel impulse response between AP and tag $n$ can be expressed as \cite{ofdm2}
\begin{eqnarray}
&\hspace{-4.8cm}\mathbf{f}_{n}=[f_{n}(1),\cdots,f_{n}(L_n^{+})]^T, \\ &\text{where } f_n (l') \sim \mathcal{CN}(0,\sigma_f^2/L_n^+), \quad
\sigma_f^2 = \frac{G_s A_e}{4\pi d^2}, \label{eqn:friis}
\end{eqnarray}
where $L_n^+$ denotes the number of multipath for tag $n \in \{1,2,\cdots,N\}$. In \eqref{eqn:friis}, $G_s$, $A_e$ and $d$, respectively, denote antenna gain at AP, effective aperture of receive antenna and tag distance from AP \cite{hybridB}. Moreover, we assume independent and identically distributed (i.i.d.) channels among tags \cite{dyadic1, dyadic2, dyadic4, dyadic5}. It is noteworthy that the forward channel with wide-band ambient signals (e.g., 20 MHz in Wi-Fi) generally experiences frequency-selective fading in sharp contrast to frequency-flat fading in RFID and bistatic backscatter, causing delay spreads of received signals in the forward channel $\mathbf{f}_{n}$ \cite{backfi}. To capture this characteristic, we can use Toeplitz matrix \cite{Toeplitz}, matrix-vector representations of circular-convolution operations. We define Toeplitz lower- and upper-triangular matrices as follows \cite{ofdm2}
\begin{eqnarray}
\mathbf{F}^{+}_n&\hspace{-0.1in}=\hspace{-0.1in}&\sum_{l'=1}^{L_n^+} f_n(l') (\mathbf{J}^{+}_L)^{l'-1} \in \mathbb{C}^{L\times L}, \\
\mathbf{F}^{-}_n&\hspace{-0.1in}=\hspace{-0.1in}&\sum_{l'=1}^{L_n^+} f_n(l') (\mathbf{J}^{-}_L)^{L-l'+1} \in \mathbb{C}^{L\times L}, \label{eqn:forward}
\end{eqnarray}
where the matrices $(\mathbf{J}^{+}_L)\in \mathbb{R}^{L \times L}$ and $(\mathbf{J}^{-}_L)\in \mathbb{R}^{L \times L}$, respectively, denote the Toeplitz forward shift and backward shift matrices of size $L \times L$ \cite{ofdm2, Toeplitz}. Due to delay spread, the received signal at tag $n$ at time slot $k$ can be expressed by superposition of $k$-th signals and $(k-1)$-th signals such that
\begin{eqnarray}
\mathbf{y}_{k,n}= \underbrace{\mathbf{F}^{+}_n \mathbf{s}_k}_{k\text{-th time}}+\underbrace{\mathbf{F}^{-}_n \mathbf{s}_{k-1}}_{k-1\text{-th time}} \in \mathbb{C}^{L\times 1}, \label{eqn:ISI}
\end{eqnarray}
where the second term in \eqref{eqn:ISI} constitutes ISI for the ambient signal $\mathbf{s}_k$.
Since tag relies on RF energy harvesting, it operates in active or idle state depending on the incident power \cite{ABCN-kaibin} of the received signal in \eqref{eqn:ISI}, which is expressed as
\begin{eqnarray}
E_n=\mathbb{E} \left [\sum_{k=1}^{K} \left \| \mathbf{y}_{k,n} \right \|^2 \right ]. \label{eqn:incident}
\end{eqnarray}
\begin{figure}
\centering
\includegraphics[width=8.9 cm]{Duty_Cycle}
\caption{Example of duty-cycling operation per codeword of length $\widetilde{K}=KL$.}
\label{fig:duty}
\end{figure}
As described in Fig.~\ref{fig:duty}, the incident power in \eqref{eqn:incident} affects duty-cycling operation and thus sparsity of NOMA. Let $0<\alpha\leq 1$ be a power ratio between backscattered and incident power, affecting performance of backscatter detection as well as energy harvesting \cite{16qam,qam,codedqam}.
Let $K_1<K$ be the number of time slots when tag is activated. In a codeword of length $\widetilde{K}=KL$, each tag is activated for $K_1 L$ samples and remains idle for $(K-K_1)L$. It is remarkable that the delay spread of backscatter channel incurs ISI for $\widetilde{L} \leq L$ samples. Moreover, backscatter symbols from $N \geq K$ tags (i.e., NOMA) in active states can be interfered in time-domain causing MAI, further complicating reliable detection of AmBC.
% Circuit-power constraint
Since RF tags are powered by ambient energy harvesting, circuit-power constraint that harvested power with efficiency $\eta$ \cite{EH-eff} is greater than circuit power for entire $\widetilde{K}$ sampling period should be satisfied. If we assume a linear power consumption model with respect to energy per symbol $\sigma_c^2$ and backscatter symbol rate $R_s$ \cite{qam, 16qam}, the circuit-power constraint can be formulated as
\begin{eqnarray}
\underbrace{D(1-\sigma_b^2)\eta E_n}_{\text{active state}}+\underbrace{(1-D)\eta E_n}_{\text{idle state}} \geq P_c,
\end{eqnarray}
where $D = \frac{K_1}{K}$, $P_c = R_s \sigma_c^2$ and $R_s = \frac{1}{L T_s}$. The average backscatter symbol power proportional to $\alpha$ is denoted by $\sigma_b^2$ and will be discussed in Section \ref{sec:enc}. For convenient analysis of EH, we define the normalized energy per backscatter symbol as
\begin{eqnarray}
\beta = D\sigma_b^2. \label{eqn:NEPS}
\end{eqnarray}
Under these settings, energy harvesting (EH) constraint or the circuit-power constraint can be reformulated as \cite{ABCN-kaibin}
\begin{eqnarray}
E_n \geq \theta = \frac{P_c}{\eta (1-\beta)}. \label{eqn:EH-constraint}
\end{eqnarray}
From \eqref{eqn:EH-constraint}, the state of tag $\mathbf{a}=[a_1,\cdots,a_N]$ (0: idle, 1: active) with EH threshold $\theta$ can be obtained by
\begin{eqnarray}
a_n &\hspace{-0.1in}=\hspace{-0.1in}& \mathbb{I}\{E_n \geq \theta\} , \quad \text{for } n \in \{1, \cdots, N\},
\end{eqnarray}
where $\mathbb{I} \{x \}$ denotes the indicator function of the event $x$.
If the EH constraint is satisfied, tag can encode sparse codewords using simple modulator circuit. In the following sections, we describe AmBC modulator in Section \ref{sec:enc} and detector in Section \ref{sec:dec}.
\section{Sparse-Coded Load Modulation} \label{sec:enc}
\begin{figure}
\centering
\includegraphics[width=8.9 cm]{Signal_Constellation}
\caption{Example of mapping function $\psi_M$ for AmBC. (a) PSK-based backscatter modulator (b) Sparse-coded backscatter modulator}
\label{fig:signal}
\end{figure}
As described in the previous section, we design AmBC modulator based on load impedances satisfying the following characteristics:
\begin{itemize}
\item {\bf Small-form Factor}: The size of the RF tags is very limited for IoT applications such as implanted biosensors and wearables, so they should have small number of load impedances to reduce their size. The proposed SC-BMA can enable $M$-ary modulation \cite{iter-Mary,code-project} for these small-form factor IoT devices, which is in sharp contrast to traditional $M$-ary backscatter modulators with massive number of load impedances (e.g., $M$-PSK \cite{backfi}, $M$-QAM \cite{16qam, mqam1, mqam2}).
\item {\bf Interference Exploitation}: Detection performance of AmBC is severely degraded by the interferences (i.e., MAI and ISI). To rather exploit these interferences for massive connectivity, we design the codeword structure capturing sparsity \cite{scma3} of ambient EH and characteristics of DBC \cite{dyadic1, ofdm2} where two-way fading and severe ISI occur.
\end{itemize}
Towad these goals, we describe the modulation and encoding models \cite{scma1,scma2,scma3,scma4}, which include design of mapping function \cite{iter-Mary,code-project} and sparse factor graph \cite{scma3}.
If tag $n$ is activated, the tag generates $\log_2 M$-length bit stream $\mathbf{d}_n^{\dagger}$ as follows
\begin{eqnarray}
\hspace{-0.6cm}\mathbf{d}_n^{\dagger} &\hspace{-0.1in}
\in\hspace{-0.1in}& \{\mathbf{d}^{\star}(1),\cdots, \mathbf{d}^{\star}(M) \}, \nonumber
\\ \hspace{-0.6cm}\mathbf{d}^{\star}(m)&\hspace{-0.1in}=\hspace{-0.1in}&[d^{\star}_1(m) \cdots d^{\star}_{\log_2 M}(m)] \in \mathbb{B}^{1 \times \log_2 M}, \quad \text{if } a_n = 1.
\end{eqnarray}
To modulate the binary vector $\mathbf{d}_n^{\dagger}$ into $M$-ary backscatter symbol vector $\mathbf{b}_n$, we define the mapping function $\psi_M$ for $K_1=2$ and $M \in \{2,4,8\}$ as
% mapping function
\begin{eqnarray}
&\begin{matrix}
\psi_2 =
\begin{cases}
\sqrt{\alpha}[1, -1], \quad &\hspace{0.01in}\text{for } \mathbf{d}_n^{\dagger}=[0], \\
\sqrt{\alpha}[-1, 1], \quad &\hspace{0.01in}\text{for } \mathbf{d}_n^{\dagger}=[1],
\end{cases} \\
\hspace{0.07in}\psi_4 =
\begin{cases}
\sqrt{\alpha}[1, 0], \quad &\text{for } \mathbf{d}_n^{\dagger}=[00], \\
\sqrt{\alpha}[0, 1], \quad &\text{for } \mathbf{d}_n^{\dagger}=[01], \\
\sqrt{\alpha}[0, -1], \quad &\text{for } \mathbf{d}_n^{\dagger}=[10], \\
\sqrt{\alpha}[-1, 0], \quad &\text{for } \mathbf{d}_n^{\dagger}=[11],
\end{cases}
\end{matrix}
\nonumber\\ &\hspace{0.6cm}\psi_8 =
\begin{cases}
\sqrt{\alpha}[1, 1], \quad &\text{for } \mathbf{d}_n^{\dagger}=[000], \\
\sqrt{\alpha}[1, 0], \quad &\text{for } \mathbf{d}_n^{\dagger}=[001], \\
\sqrt{\alpha}[0, -1], \quad &\text{for } \mathbf{d}_n^{\dagger}=[010], \\
\sqrt{\alpha}[1, -1], \quad &\text{for } \mathbf{d}_n^{\dagger}=[011], \\
\sqrt{\alpha}[0, 1], \quad &\text{for } \mathbf{d}_n^{\dagger}=[100], \\
\sqrt{\alpha}[-1, 1], \quad &\text{for } \mathbf{d}_n^{\dagger}=[101], \\
\sqrt{\alpha}[-1, -1], \quad &\text{for } \mathbf{d}_n^{\dagger}=[110], \\
\sqrt{\alpha}[-1, 0], \quad &\text{for } \mathbf{d}_n^{\dagger}=[111].
\end{cases}
\label{eqn:mapping}
\end{eqnarray}
Then, the $M$-ary binary vector $\mathbf{b}_n$ can be expressed by
\begin{eqnarray}
\mathbf{b}_n &\hspace{-0.1in}=\hspace{-0.1in}& \psi_M(\mathbf{d}_n^{\dagger}) \in \{\mathbf{b}^{\star}(1), \cdots, \mathbf{b}^{\star}(M) \},
\\
\mathbf{b}^{\star}(m) &\hspace{-0.1in}=\hspace{-0.1in}& [b_1^{\star}(m), \cdots, b_{K_1}^{\star}(m)] \in \mathbb{R}^{1 \times K_1}, \nonumber\\
\sigma_b^2 &\hspace{-0.1in}=\hspace{-0.1in}& \mathbb{E}\left [ \frac{\|\mathbf{b}_n \|^2}{K_1}\right ]=
\begin{cases}
\alpha, \quad &\text{for } M = 2, \\
0.5\alpha, \quad &\text{for } M = 4,\\
0.75 \alpha, \quad &\text{for } M = 8.
\end{cases} \label{eqn:backpower}
\end{eqnarray}
As described in Fig.~\ref{fig:signal}, these mapping functions satisfy the small-form factor requirement in RF tags. For instance, in conventional phase-shift keying (PSK)-based backscatter modulator, $M$ constellation points are present in signal constellation with phase difference $2\pi/M$ in Fig.~\ref{fig:signal}. (a) \cite{backfi}.
So the modulator requires $N_Z = M$ distinct load impedances, inevitably increasing size of RF tags.
On the other hand, SC-AmBC modulator only has 3 constellation points even for higher-order modulation. Accordingly, the required number of load impedances is $N_Z=2$, since zero-constellation point is equivalent to the idle state of RF tag \cite{codingB}. Although several constellation points can be overlapped in Fig.~\ref{fig:signal}. (b), symbols can be detectable as long as they are uniquely mapped in $K_1$-dimensional signal space in \eqref{eqn:mapping}, even achieving {\em modulation diversity gain} \cite{mod-div} for reliable AmBC.
Thus, high-order modulation can be implemented to typical binary PSK (BPSK) backscatter modulator, enabling high data rate for small-form factor passive RF devices such as biosensors and wearables.
In the sequel, $K_1$-length modulated symbols $\mathbf{b}_n$ are spread into $K$-length sparse codewords using one-way factor graph (Tanner graph) $\widetilde{\mathbf{G}}$ \cite{LDS}.
By doing so, dimension of signal is extended to $2K$-dimension \cite{scma1, scma2, scma3, scma4} with $M^N$ points in main constellation, in contrast to conventional AmBC modulators typically laid on 2-dimensional Euclidean spaces \cite{wireless} with $M$ constellation points (e.g., 4-QAM \cite{qam,codedqam}, 16-QAM \cite{16qam,mqam1}, 16-PSK \cite{backfi}, 32/64-QAM \cite{mqam2}).
As a result, the number of tags allowed to transmit is increased to $N=\binom{K}{K_1}$. The overloading factor \cite{scma1}, ratio of the number of RF tags and the time slots can be defined by
\begin{eqnarray}
\lambda = N/K \geq 1.
\end{eqnarray}
In addition, the duty cycle of tag is extended to $D=K_1/K$, which is $K_1$-fold increase compared to the time-division-based AmBC (TD-AmBC) with $D=1/K$ \cite{backfi, ambB-TD}. Although $N_1=\binom{K-1}{K_1-1}$ tags are interfering in arbitrary time slot, the MAI can be efficiently suppressed by exploiting sparse code structure. We choose $K_1=2$ to achieve maximum sparsity as well as reasonable error rate for backscatter detection \cite{scma3}.
\begin{figure}
\centering
\includegraphics[width=8.9 cm]{Ex_Factor_Graph}
\caption{Example of dyadic factor graph composed of forward factor graph $\widetilde{\mathbf{G}}^{+}$ and backward factor graph $\widetilde{\mathbf{G}}^{-}$ ($K=4$, $N=6$, $K_1=2$, $L=3$, $\widetilde{L}=1$).}
\label{fig:Ex_FG}
\end{figure}
Note that in AmBC, to conserve energy as well as increase SNR for RF tags, $K$-length codewords are spanned on $L$ samples, generating $\widetilde{K}=KL$-length codewords. Also, ambient backscatter signal experiences DBC in contrast to the one-way Rayleigh channel in conventional sparse coding. Based on these observations, we introduce dyadic factor graph extending one-way factor graph $\widetilde{\mathbf{G}}$ as shown in Fig.~\ref{fig:Ex_FG}. The dyadic (two-way) factor graph has 2 kinds of nodes, namely, variable node (VN) representing codeword of tag $n \in \{1,\cdots,N\}$ and function node (FN) representing received signal at sample $\widetilde{k} \in \{1,\cdots,\widetilde{K}\}$.
Due to ISI among backscatter signals, the dyadic factor graph is composed of the forward factor graph $\widetilde{\mathbf{G}}^{+} =\{\widetilde{g}^{+}_{\widetilde{k},n}\} \in \mathcal{B}^{\widetilde{K} \times N}$ and the backward factor graph $\widetilde{\mathbf{G}}^{-} =\{\widetilde{g}^{-}_{\widetilde{k},n}\}\in \mathcal{B}^{\widetilde{K} \times N}$ representing interference symbols. Here, we use the superscript $\{\cdot\}^-$ to represent the signals related to ISI and the superscript $\{\cdot\}^+$ vice versa.
The backward factor graph has $K\widetilde{L}$ non-zero rows which are circular-shifted from the forward factor graph, and sparse matrix under the condition $\widetilde{L}<L$, stating that symbols are only interfered by neighboring symbols \cite{ofdm2}.
The new graph structure can efficiently exploit the MAI and ISI for backscatter detection, satisfying the second requirement of AmBC for massive connectivity. Based on sparse factor graph structure \cite{scma3} in UL, the backscatter modulation can be generalized by the pseudo codes in Algorithm \ref{alg:code}.
\begin{algorithm}
\caption{Sparse-coded backscatter modulation algorithm (SC-BMA)}
\begin{algorithmic}[1]
\State {\bf input } the number of time slots $K$ and the modulation order $M$
\State obtain the symbol mapping $\mathbf{b}^{\star}(m)=\psi_M(\mathbf{d}^{\star}(m)), \quad m \in \{1,\cdots,M\}$ from \eqref{eqn:mapping}
\For{$m \in \{1,\cdots,M\}$}
\State {\bf initialize} $n=1$
\For{$n_1=1:K-1$}
\For{$n_2=1:K-n_1$}
\For{$k=1:K$}
\If{$k=n_1$}
\State $\widetilde{b}_{k,n}(m) = b_1^{\star}(m), \quad \widetilde{g}_{k,n} = 1$ \Comment{first non-zero element}
\ElsIf{$k=n_1+n_2$}
\State $\widetilde{b}_{k,n}(m) = b_2^{\star}(m), \quad \widetilde{g}_{k,n} = 1$ \Comment{second non-zero element}
\Else
\State $\widetilde{b}_{k,n}(m) = 0, \quad \widetilde{g}_{k,n} = 0$ \Comment{zero element}
\EndIf
\EndFor
\State $n \leftarrow n+1$
\EndFor
\EndFor
\EndFor
\State obtain $\widetilde{\mathbf{G}} = \{\widetilde{g}_{k,n}\}$, $\mathbf{B}(m) = \{\widetilde{b}_{k,n}(m)\}$ \Comment{one-way factor graph, codebook}
\State generate $\widetilde{\mathbf{G}}^{+} = \widetilde{\mathbf{G}}\otimes \mathbf{1}_L$, $\mathbf{B}^{+}(m) = \mathbf{B}(m)\otimes \mathbf{1}_L$
\State generate $\widetilde{\mathbf{G}}^{-} =[(\mathbf{J}_K^{+}+(\mathbf{J}_K^{-})^{K-1})\widetilde{\mathbf{G}}] \otimes [\mathbf{1}_{\widetilde{L}}^T ~\mathbf{0}_{L-\widetilde{L}}^T]^T$, $\mathbf{B}^{-}(m)=[(\mathbf{J}_K^{+}+(\mathbf{J}_K^{-})^{K-1})\mathbf{B}(m)] \otimes [\mathbf{1}_{\widetilde{L}}^T ~\mathbf{0}_{L-\widetilde{L}}^T]^T$
\State {\bf output} the codebooks $\{\mathbf{B}^{+}(m), \mathbf{B}^{-}(m)\}_{m \in \{1,\cdots,M\}}$ and the factor graphs $\widetilde{\mathbf{G}}^{+}, \widetilde{\mathbf{G}}^{-}$
\end{algorithmic} \label{alg:code}
\end{algorithm}
%
Since channel information is not available at RF tags, they encode data using the one-way codebook $\{\mathbf{B}(m)\}_{m \in \{1,\cdots,M\}}$ without knowing ISI, and the forward and backward codebooks $\{\mathbf{B}^{+}(m), \mathbf{B}^{-}(m)\}_{m \in \{1,\cdots,M\}}$ containing information on ISI will be used in detection procedure in Section \ref{sec:dec}. So the reflection coefficients and backscattered signals at RF tags over $L$ samples can be represented by
\begin{eqnarray}
&\mathbf{x}_{k,n}=\Gamma_{k,n}\mathbf{y}_{k,n}\in \mathbb{C}^{1\times L}, \nonumber\\ &\text{where }
\Gamma_{k,n} \in
\begin{cases}
\{0\},\quad &\text{if } a_n=0 \text{ (idle)}, \\
\widetilde{b}_{k,n}(m), \quad &\text{if } a_n=1 \text{ (active)}.
\end{cases} \label{eqn:gamma}
\end{eqnarray}
Similar to the forward channel, backward channel and interference channel can be defined by
\begin{eqnarray}
\mathbf{g}_{n}&\hspace{-0.1in}=\hspace{-0.1in}&[g_{n}(1),\cdots,g_{n}(L_n^-)]^T, \quad
\mathbf{h}_0=[h_0(1),\cdots,h_0(L_0)]^T, \nonumber \\
\mathbf{G}^{+}_n&\hspace{-0.1in}=\hspace{-0.1in}&\sum_{l'=1}^{L_n^-} g_n(l') (\mathbf{J}^{+}_L)^{l'-1} \in \mathbb{C}^{L\times L}, \nonumber\\
\mathbf{G}^{-}_n&\hspace{-0.1in}=\hspace{-0.1in}&\sum_{l'=1}^{L_n^-} g_n(l') (\mathbf{J}^{-}_L)^{L-l'+1} \in \mathbb{C}^{L\times L}, \nonumber \\
\mathbf{H}^{+}_0&\hspace{-0.1in}=\hspace{-0.1in}&\sum_{l'=1}^{L_0} h_0(l') (\mathbf{J}^{+}_L)^{l'-1} \in \mathbb{C}^{L\times L}, \nonumber\\
\mathbf{H}^{-}_0&\hspace{-0.1in}=\hspace{-0.1in}&\sum_{l'=1}^{L_0} h_0(l') (\mathbf{J}^{-}_L)^{L-l'+1} \in \mathbb{C}^{L\times L}, \label{eqn:backch}
\end{eqnarray}
where $g_n (l') \sim \mathcal{CN}(0,\sigma_g^2/L_n^-)$, $h_0 (l') \sim \mathcal{CN}(0,\sigma_{h_0}^2/L_0)$ and the variance of backward and leakage channels are denoted by $\sigma_g^2$ and $\sigma_{h_0}^2$, respectively. In \eqref{eqn:backch}, $L_n^-$ and $L_0$ denote the number of multipaths for backward channel and leakage channel, respectively.
Then, the signal received at AP at time slot $k$ can be expressed as
\begin{eqnarray}
\mathbf{r}_k &\hspace{-0.1in}=\hspace{-0.1in}& (\mathbf{H}_0^{+}\mathbf{s}_k + \mathbf{H}_0^{-}\mathbf{s}_{k-1})+
\sum_{n=1}^N (\mathbf{G}_n^{+}\mathbf{x}_{k,n}+\mathbf{G}_n^{-}\mathbf{x}_{k-1,n})+\mathbf{w}_k \nonumber \\
%
&\hspace{-0.1in}=\hspace{-0.1in}&\sum_{n=1}^N \underbrace{(\mathbf{G}_n^{+}\mathbf{F}_n^{+}\mathbf{s}_{k}+
\mathbf{G}_n^{+}\mathbf{F}_n^{-}\mathbf{s}_{k-1})}_{=~\widetilde{\mathbf{h}}^{+}_{k,n}}\Gamma_{k,n}
\nonumber\\
\quad &\hspace{-0.1in}{}\hspace{-0.1in}&+\sum_{n=1}^N \underbrace{(\mathbf{G}_n^{-}\mathbf{F}_n^{+}\mathbf{s}_{k-1}+\mathbf{G}_n^{-}\mathbf{F}_n^{-}\mathbf{s}_{k-2})}_{=~\widetilde{\mathbf{h}}^{-}_{k,n}}\Gamma_{k-1,n}
\nonumber \\
%
\quad &\hspace{-0.1in}{}\hspace{-0.1in}&+ \underbrace{(\mathbf{H}_0^{+}\mathbf{s}_k+ \mathbf{H}_0^{-}\mathbf{s}_{k-1})}_{=~\widetilde{\mathbf{h}}_{k,0}}
+\mathbf{w}_k \nonumber \\
%
&\hspace{-0.1in}=\hspace{-0.1in}& \underbrace{\sum_{n=1}^N \widetilde{\mathbf{h}}^{+}_{k,n}\Gamma_{k,n}}_{\text{forward codeword (MAI)}}+
\underbrace{\sum_{n=1}^N \widetilde{\mathbf{h}}^{-}_{k,n}\Gamma_{k-1,n}}_{\text{backward codeword (ISI)}}\nonumber \\
\quad &\hspace{-0.1in}{}\hspace{-0.1in}&+\underbrace{\widetilde{\mathbf{h}}_{k,0}}_{\text{SI}}+\underbrace{\mathbf{w}_k}_{\text{noise}}
\in \mathbb{C}^{L \times 1}, \label{eqn:superposition}
\end{eqnarray}
where $\mathbf{w}_k = [w_{1,k},\cdots,w_{L,k}]^T$ denotes additive white Gaussian noise with variance $\sigma_n^2$. Since the AP knows the ambient signal $\mathbf{s}_k$ for $k \in \{1,\cdots, K\}$ and the leakage channel $\mathbf{h}_0$, the SI term in \eqref{eqn:superposition} can be reconstructed for SIC \cite{backfi}. The SIC ouput signal after hybrid digital and analog cancellation \cite{SIC} can be given by
\begin{eqnarray}
\mathbf{t}_k = \mathbf{r}_k-\widetilde{\mathbf{h}}_{k,0} \in \mathbb{C}^{L \times 1}. \label{eqn:SIC}
\end{eqnarray}
However, as discussed in \cite{backfi,ofdm2}, AP estimates the composite forward-backward channel $\{\mathbf{h}_n\}_{n \in \{1,\cdots,N\}}$ rather than the individual estimates of the forward channels $\{\mathbf{f}_n\}_{n \in \{1,\cdots,N\}}$ and the backward channels $\{\mathbf{g}_n\}_{n \in \{1,\cdots,N\}}$. The channels $\mathbf{h}_n$ can be efficiently obtained by compressed sensing based channel estimation techniques \cite{cs-survey}, which can be defined as
\begin{eqnarray}
\mathbf{h}_{n}&\hspace{-0.1in}=\hspace{-0.1in}&[h_{n}(1),\cdots,h_{n}(L_n^\star)]^T, \nonumber\\
\mathbf{H}^{+}_n&\hspace{-0.1in}=\hspace{-0.1in}&\sum_{l'=1}^{L_n^\star} h_n(l') (\mathbf{J}^{+}_L)^{l'-1} \in \mathbb{C}^{L\times L}, \nonumber\\
\mathbf{H}^{-}_n&\hspace{-0.1in}=\hspace{-0.1in}&\sum_{l'=1}^{L_n^\star} h_n(l') (\mathbf{J}^{-}_L)^{L-l'+1} \in \mathbb{C}^{L\times L},
\end{eqnarray}
where $L_n^\star = L_n^{+}+L_n^{-}-1$ denotes the number of multipaths for the composite channel.
Assuming i.i.d. fully-correlated DBC model with channel reciprocity appearing in in-band wireless-powered communication network \cite{wpcn}, the backward channel satisfies $\mathbf{g}_n = \mathbf{f}_n$ and has $L^{-} = L_n^{-} =L^{+}= L_n^{+}$ multipaths. Consequently, for AmBC detection, we should estimate the forward channels with length $L^{+}$ from the dyadic channels with $L^\star = 2 L^{+}-1$. Based on the circular convolution operation $\ast$, the relation between $\mathbf{f}_n$ and $\mathbf{h}_n$ can be represented by
\begin{eqnarray}
\hspace{-0.6cm} \mathbf{h}_n &\hspace{-0.1in}=\hspace{-0.1in}& \mathbf{f}_n \ast \mathbf{f}_n, \nonumber\\
\hspace{-0.6cm} h_n(l') &\hspace{-0.1in}=\hspace{-0.1in}& \sum_{i'} f_n (i') f_n (l' - i' +1), \quad l' \in \{1, \cdots, L^\star \}. \label{eqn:circ-conv}
\end{eqnarray}
For example, if $L^{+}=3$, \eqref{eqn:circ-conv} can be expressed by
\begin{eqnarray}
h_n(1) &\hspace{-0.1in}=\hspace{-0.1in}& f_n(1)f_n(1)=f_n(1)^2, \label{eqn:h1}\\
h_n(2) &\hspace{-0.1in}=\hspace{-0.1in}& f_n(1)f_n(2)+f_n(2)f_n(1) = 2f_n(1)f_n(2), \label{eqn:h2} \\
h_n(3) &\hspace{-0.1in}=\hspace{-0.1in}& f_n(1)f_n(3)+f_n(2)f_n(2)+f_n(3)f_n(1) \nonumber \\&\hspace{-0.1in}=\hspace{-0.1in}&f_n(2)^2+2f_n(1)f_n(3), \label{eqn:h3}\\
h_n(4) &\hspace{-0.1in}=\hspace{-0.1in}& f_n(2)f_n(3)+f_n(3)f_n(2) = 2f_n(2)f_n(3), \\
h_n(5) &\hspace{-0.1in}=\hspace{-0.1in}& f_n(3)f_n(3) = f_n(3)^2.
\end{eqnarray}
From \eqref{eqn:h1}, $f_n(1)$ can be given by $f_n(1) = \pm \sqrt{h_n(1)}$. we choose the estimated channel $\hat{f}_n(1) = \sqrt{h_n(1)}$ from \eqref{eqn:h1} and then sequentially solve \eqref{eqn:h2} and \eqref{eqn:h3}. From \eqref{eqn:superposition}, by solving the first $L^{+}$ equations, the vector $\hat{\mathbf{f}}_n$ and the corresponding matrices $\hat{\mathbf{F}}_n^{+}$ and $\hat{\mathbf{F}}_n^{-}$ can be obtained from \eqref{eqn:forward}. Then, the dyadic channels in \eqref{eqn:superposition} can be reformulated as
\begin{eqnarray}
\widetilde{\mathbf{h}}^{+}_{k,n} \left(\hat{\mathbf{f}}_n\right)&\hspace{-0.1in}=\hspace{-0.1in}&\hat{\mathbf{F}}_n^{+}\hat{\mathbf{F}}_n^{+}\mathbf{s}_k +\hat{\mathbf{F}}_n^{+}\hat{\mathbf{F}}_n^{-}\mathbf{s}_{k-1}, \label{eqn:comp1}\\
\widetilde{\mathbf{h}}^{-}_{k,n}\left(\hat{\mathbf{f}}_n\right)&\hspace{-0.1in}=\hspace{-0.1in}&\hat{\mathbf{F}}_n^{-}\hat{\mathbf{F}}_n^{+}\mathbf{s}_{k-1} , \label{eqn:comp2}
\end{eqnarray}
where the second term $\hat{\mathbf{F}}_n^{-}\hat{\mathbf{F}}_n^{-}\mathbf{s}_{k-2}=\mathbf{0}_{L}$ is neglected under the condition $L^\star \leq L$ \cite{ofdm2}.
\\ {\bf Remark 1}
The dyadic channels are even such that
\begin{eqnarray}
\widetilde{\mathbf{h}}^{+}_{k,n} \left(\hat{\mathbf{f}}_n\right)&\hspace{-0.1in}=\hspace{-0.1in}&\widetilde{\mathbf{h}}^{+}_{k,n} \left(-\hat{\mathbf{f}}_n\right), \\
\widetilde{\mathbf{h}}^{-}_{k,n} \left(\hat{\mathbf{f}}_n\right)&\hspace{-0.1in}=\hspace{-0.1in}&\widetilde{\mathbf{h}}^{-}_{k,n} \left(-\hat{\mathbf{f}}_n\right).
\end{eqnarray}
Since $\hat{\mathbf{f}}_n = \pm \mathbf{f}_n$, we may incorrectly estimate the channels when $\hat{\mathbf{f}}_n = - \mathbf{f}_n$. However, the even symmetry of the channels ensures correct estimation of the channels regardless of their directions. It is noteworthy that the previous works \cite{backfi,ofdm2} in AmBC assumed $\widetilde{\mathbf{h}}^{+}_{k,n}=\mathbf{H}_n^{+}\mathbf{s}_k$ and $\widetilde{\mathbf{h}}^{-}_{k,n}=\mathbf{0}_L$ by simply discarding $L^{+}-1$ samples affected by ISI, inevitably leading to SNR reduction by factor $1-(L^{+}-1)/L$. On the other hand, the proposed channel estimation based on the Algorithm \ref{alg:estim} enables exploitation of these interferences without SNR loss.
In the following section, we discuss iterative detection scheme for superposition signals using dyadic channels as well as codebooks.
\begin{algorithm}
\caption{Dyadic channel estimation algorithm (D-CEA)}
\begin{algorithmic}[1]
\State {\bf input } the composite forward-backward channels $\left\{\mathbf{h}_n\right\}_{n \in \{1,\cdots,N\}}$ and the ambient source signals $\left\{\mathbf{s}_k\right\}_{k \in \{1,\cdots,K\}}$
\For {$n \in \{1, \cdots, N\}$}
\State {\bf initialize } $f_n(1) = \sqrt{h_n(1)}$
\For {$l' = 2:(L^\star+1)/2$}
\State solve \eqref{eqn:circ-conv} for the variable $f_n(l')$
\EndFor
\State estimate the channel vector $\hat{\mathbf{f}}_n=[\hat{f}_n(1),\cdots,\hat{f}_n((L^\star+1)/2)]^T$
\State estimate the Toeplitz matrices $\hat{\mathbf{F}}_n^{+}$ and $\hat{\mathbf{F}}_n^{-}$ from the equation \eqref{eqn:forward}
\For {$k \in \{1,\cdots,K\}$}
\State estimate the dyadic forward and backward channels $\widetilde{\mathbf{h}}^{+}_{k,n}$ and $\widetilde{\mathbf{h}}^{-}_{k,n}$ from \eqref{eqn:comp1}, \eqref{eqn:comp2}
\EndFor
\EndFor
\State {\bf output } the dyadic foward channel $\left\{\widetilde{\mathbf{h}}^{+}_{k,n}\right\}_{\forall k,n}$ and the dyadic backward channel $\left\{\widetilde{\mathbf{h}}^{-}_{k,n}\right\}_{\forall k,n}$
\end{algorithmic} \label{alg:estim}
\end{algorithm}
\begin{figure}
\centering
\includegraphics[width=8.9 cm]{D-MPA}
\caption{Example of AmBC signals in DBC ($K=4$, $L=8$, $L^{+}=3$). (a) Transmitted signal from AP (b) Received signal at tag with ISI from forward channel (c) Superposition signal at AP with ISI from dyadic channel}
\label{fig:D-MPA}
\end{figure}
\section{Iterative Detection on Dyadic Channel} \label{sec:dec}
For reliable detection of AmBC on DBC, we design the iterative detector with the following characteristics:
\begin{itemize}
\item {\bf Interference Exploitation}: To achieve massive connectivity in AmBC, the interferences (i.e., MAI and ISI) should be exploited in non-orthogonal manners \cite{scma1, FBMC}, instead of excluding them by using orthogonal signaling as in conventional detectors \cite{backfi, ofdm2}. In addition, iterative MPA should be modified for DBC to achieve additional gain without loss of backscatter signal power.
\item {\bf Low Complexity}: The received signals in AP are superposition of multiple backscatter signals with $M$-ary modulation. Also, ISI introduces another superposition of backscatter signals with delay spreads in the dyadic channels, further increasing complexity of iterative detector. Hence, the complexity of detector should be as low as possible for practical applications of SC-AmBC.
\end{itemize}
To achieve these goals, we describe iterative detector especially on DBC in this section. The SIC output \cite{SIC} in \eqref{eqn:SIC} is rewritten by superposition of $N$ reflected signals in time slot $k$ and delayed signals in time slot $k-1$, which is
\begin{eqnarray}
\mathbf{t}_k &\hspace{-0.1in}=\hspace{-0.1in}&[t_{1,k},\cdots,t_{L,k}]^T \nonumber\\
&\hspace{-0.1in}=\hspace{-0.1in}& \sum_{n=1}^{N} \widetilde{\mathbf{h}}_{k,n}^{+} \Gamma_{k,n}\nonumber\\
&\hspace{-0.1in}+\hspace{-0.1in}&\sum_{n=1}^{N} \widetilde{\mathbf{h}}_{k-1,n}^{-} \Gamma_{k-1,n}+\mathbf{w}_k\in \mathbb{C}^{L\times 1}, \\
t_{l,k} &\hspace{-0.1in}=\hspace{-0.1in}& \sum_{n=1}^{N} \widetilde{h}_{k,n,l}^{+}\Gamma_{k,n}+\sum_{n=1}^{N} \widetilde{h}_{k,n,l}^{-}\Gamma_{k-1,n}+w_{l,k} \in \mathbb{C}. \nonumber \label{eqn:SIC-SCMA}
\end{eqnarray}
Let the $n$-th codeword $\boldsymbol{\gamma}_{n}=[\Gamma_{1,n},\cdots,\Gamma_{K,n}]^T \in \mathcal{X}_n$ where $\mathcal{X}_n \subset \mathbb{C}^{K \times 1}$ is the codebook of tag $n$ with cardinality $|\mathcal{X}_n|=M$.
Define a codeword matrix $\mathbf{X}=[\boldsymbol{\gamma}_{1},\cdots,\boldsymbol{\gamma}_{N}] \in \mathbb{C}^{K \times N}$ and a received signal matrix $\mathbf{Y}=[\mathbf{t}_1,\cdots,\mathbf{t}_K] \in \mathbb{C}^{K \times N}$.
Then the maximum {\em a posteriori} (MAP) detection rule for codeword $\mathbf{X}$ can be expressed as \cite{scma1}
\begin{eqnarray}
\hat{\mathbf{X}} = \arg \max_{\mathbf{X}\in \mathcal{X}_1 \times\cdots \times \mathcal{X}_N} p ( \mathbf{X} | \mathbf{Y} ). \label{eqn:MAP}
\end{eqnarray}
The complexity of the above problem is shown to increases exponentially with the number of tags $N$ and polynomially with the constellation size $M$ \cite{scma1}. Fortunately, the MAP detection problem in \eqref{eqn:MAP} can be translated into marginalize product of function (MPF) problem which can be efficiently solved by MPA \cite{LDS}.
In MPA, where codewords are iteratively decoded over the factor graph, we can achieve near-optimal detection performance with low complexity \cite{scma1, scma2, scma3, scma4}.
However, as described in Fig.~\ref{fig:D-MPA}, DBC incurs significant ISI for detection of codeword $\mathbf{X}$. If we simply filter ISI using inaccurate channel estimation as in \cite{backfi, ofdm2}, we lose large portion of signal power (e.g., maximum 62.5\% in Fig.~\ref{fig:D-MPA}. (c)) especially for high-rate AmBC applications.
Thus, to prevent the SNR loss caused by DBC, we should modify the conventional MPA \cite{LDS} to exploit the weighted-sum of ISI.
In addition, we adopt the log-domain MPA \cite{log-MAP,LDS}, which shows superior decoding performance than MPA for reliable detection of AmBC at low SNR. To achieve this goal, we define the dyadic factor graph for ambient backscatter as follows:
\begin{eqnarray}
\widetilde{\mathbf{G}}^\star =\widetilde{\mathbf{G}}^{+} \oplus \widetilde{\mathbf{G}}^{-},
\end{eqnarray}
where the operator $\oplus$ denotes element-wise OR operation. In addition, by the channel reciprocity, $\widetilde{L} = L^{+}-1$.
\\{\bf Remark 2}
The dyadic factor graph $\widetilde{\mathbf{G}}^\star$ with $\widetilde{K}=KL$ FNs and $N$ VNs satisfies the following properties:
\begin{itemize}
\item There are $K(L-\widetilde{L})$ FNs with degree $K-1$ and $K\widetilde{L}$ FNs with degree $2K-3$.
\item There are $K$ VNs with degree $2L+\widetilde{L}$ and $N-K$ VNs with degree $2L+2\widetilde{L}$.
\end{itemize}
If $\widetilde{L}=0$, the graph is equivalent to the spreading of the one-way factor graph $\widetilde{\mathbf{G}}$ with spreading factor $L$. As described in Fig. \ref{fig:Ex_FG}, $\widetilde{L}$ samples are collided by previous backscatter symbols causing ISI and increasing complexity of decoding. The condition $\widetilde{L} < L$ ensures low-complexity decoding since the collision is only affected by neighboring backscatter symbols \cite{ofdm2}. Especially, from Remark 2, information exchanged between FNs and VNs can be decomposed into two cases for reducing computational complexity while exploiting ISI. To perform decoding, we arrange the received signals and channels as follows
\begin{eqnarray}
\vec{\mathbf{y}}&\hspace{-0.1in}=\hspace{-0.1in}&[\vec{y}_1,\cdots,\vec{y}_{\widetilde{K}}]^T
=[\mathbf{t}_1^T, \cdots, \mathbf{t}_K^T]^T, \\
\vec{\mathbf{h}}_n^{+}&\hspace{-0.1in}=\hspace{-0.1in}&\left[\vec{h}_{1,n}^{+}, \cdots, \vec{h}_{\widetilde{K},n}^{+}\right]^T=
\left[\left(\widetilde{\mathbf{h}}^{+}_{1,n}\right)^T,\cdots,\left(\widetilde{\mathbf{h}}^{+}_{K,n}\right)^T\right]^T, \\
\vec{\mathbf{h}}_n^{-}&\hspace{-0.1in}=\hspace{-0.1in}&\left[\vec{h}_{1,n}^{-}, \cdots, \vec{h}_{\widetilde{K},n}^{-}\right]^T=
\left[\left(\widetilde{\mathbf{h}}^{-}_{1,n}\right)^T,\cdots,\left(\widetilde{\mathbf{h}}^{-}_{K,n}\right)^T\right]^T.
\end{eqnarray}
Then, we define symbol vector composed of $\widetilde{M} = \min(M,3) \leq M$ components as
\begin{eqnarray}
\vec{\mathbf{s}} =
\begin{cases}
[\vec{S}(\widetilde{m}_1), \vec{S}(\widetilde{m}_2)] = \sqrt{\alpha}[1, -1], \quad & \text{for } M = 2,\\
[\vec{S}(\widetilde{m}_1), \vec{S}(\widetilde{m}_2), \vec{S}(\widetilde{m}_3)] = \sqrt{\alpha}[1, 0, -1], \quad & \text{for } M \geq 4.
\end{cases}
\end{eqnarray}
For notational convenience, we define the ISI condition as follows
\begin{eqnarray}
\mathcal{D}(\widetilde{k},n) = \sum_{n=1}^N a_n\widetilde{g}_{\widetilde{k},n}^{-} >0.
\end{eqnarray}
Given the received signal $\vec{\mathbf{y}}$, the channels $\{\vec{\mathbf{h}}_n^{+},\vec{\mathbf{h}}_n^{-}\}_{\forall n}$, the noise variance $\sigma_n^2$, the tag state $\mathbf{a}$ \cite{rateless}, the factor graphs $\{\widetilde{\mathbf{G}}^{+}, \widetilde{\mathbf{G}}^{-}, \widetilde{\mathbf{G}}^{\star}\}$ and the codebooks $\{\mathbf{B}^{+}(m),\mathbf{B}^{-}(m)\}_{\forall m}$, the initial information on FN $\widetilde{k}$ can be formulated as
\begin{eqnarray}
&\hspace{-7cm}f^{\dagger}_{\widetilde{k}} (\{m_i\}_{i \in \mathcal{I}_{\widetilde{k}}}) \nonumber\\
&\hspace{-0.4cm}\leftarrow\begin{cases}
-\frac{1}{2\sigma_n^2} \left\|\vec{y}_{\widetilde{k}}-\sum_{i\in\mathcal{I}_{\widetilde{k}}} \left[\vec{h}_{\widetilde{k},i}^{+}B_{\widetilde{k},i}^{+}(m_{i})+ \vec{h}_{\widetilde{k},i}^{-}B_{\widetilde{k},i}^{-}(m_{i})\right]\right \|^2, \quad \\ \hspace{5cm} \text{if } \mathcal{D}(\widetilde{k},n),\label{eqn:fk_ISI}\\
-\frac{1}{2\sigma_n^2} \left\|\vec{y}_{\widetilde{k}}-\sum_{i\in\mathcal{I}_{\widetilde{k}}} \vec{h}_{\widetilde{k},i}^{+}B_{\widetilde{k},i}^{+}(m_{i})\right \|^2, \quad \\ \hspace{5cm}\text{otherwise},
\end{cases}
\label{eqn:fk}
\end{eqnarray}
where the set $\mathcal{I}_{\widetilde{k}}=\left\{n\big|a_n \widetilde{g}_{\widetilde{k},n}^{\star}>0\right\}$ denotes the VN indices connected to FN $\widetilde{k}$ and codeword indices satisfy $m_i \in \{1, \cdots, M\}$.
If FN $\widetilde{k}$ is affected by ISI, the initial information is modified by adding the weighted-sum of ISI represented by the codebook $\{\mathbf{B}^{-}(m)\}_{\forall m}$ in \eqref{eqn:fk}.
On the other hand, the dimension of the information in \eqref{eqn:fk} increases exponentially as modulation order $M$ increases \cite{iter-Mary, code-project}. Alternatively, we can utilize the fact that the number of symbols is much smaller than the number of codewords as described in Fig.~\ref{fig:signal}. For example, if $M=8$, total 8 codewords are represented in $K_1=2$ dimension with 3 overlapped symbols.
The non-orthogonal projections of the mapping function enable low-complexity decoding even for higher values of $M$ by compression of the information.
Using codeword projection method \cite{iter-Mary, code-project}, the information in \eqref{eqn:fk} can be reformulated as
\begin{eqnarray}
&\hspace{-7cm}\vec{f}_{\widetilde{k}}^{-} (\{{m}_i\}_{i \in \mathcal{I}_{\widetilde{k}}})
\nonumber \\
&\leftarrow -\frac{1}{2\sigma_n^2} \left\|\vec{y}_{\widetilde{k}}-\sum_{i\in\mathcal{I}_{\widetilde{k}}} \left[\vec{h}_{\widetilde{k},i}^{+}B_{\widetilde{k},i}^{+}(m_{i})+ \vec{h}_{\widetilde{k},i}^{-}B_{\widetilde{k},i}^{-}(m_{i})\right]\right \|^2, \nonumber\\
&\text{if } \mathcal{D}(\widetilde{k},n), \\
&\hspace{-7cm}\vec{f}_{\widetilde{k}}^{+} (\{\widetilde{m}_i\}_{i \in \mathcal{I}_{\widetilde{k}}}) \nonumber\\
&\hspace{-2.9cm}\leftarrow -\frac{1}{2\sigma_n^2} \left\|\vec{y}_{\widetilde{k}}-\sum_{i\in\mathcal{I}_{\widetilde{k}}} \vec{h}_{\widetilde{k},i}^{+}\vec{S}(\widetilde{m}_{i})\right \|^2, \nonumber \\
&\text{otherwise}. \label{eqn:nonISI}
\end{eqnarray}
Now, computational complexity at each FN is reduced for the non-ISI condition in \eqref{eqn:nonISI}.
Then, we assign {\em a priori} probability $ap_{n}(m_{n})$ to initial information from VN $n$ to FN $\widetilde{k}$ as follows:
\begin{eqnarray}
\mathrm{I}_{n\rightarrow \widetilde{k}}(m_{n})\leftarrow ap_{n}(m_{n})=\frac{1}{M}.
\end{eqnarray}
In the log-domain MPA, cascades of logarithm, summation and exponential operation are approximated using Jacobian logarithm and correction function for low-SNR applications. Now, we define a function $\overset{\star}{\max}(\delta_1,\delta_2)$ as \cite{log-MAP,LDS}
\begin{eqnarray}
\ln (e^{\delta_1}+e^{\delta_2}) \approx \overset{\star}{\max}(\delta_1,\delta_2)
&\hspace{-0.1in}=\hspace{-0.1in}&\max (\delta_1,\delta_2) \nonumber\\
&\hspace{-0.1in}+\hspace{-0.1in}&\ln (1+e^{-|\delta_2-\delta_1|}).
\end{eqnarray}
In the next step, FN $\widetilde{k}$ passes the updated messages to its neighboring VNs. The intermediate variable for the log-domain MPA can be calculated as
\begin{eqnarray}
&\hspace{-6.5cm}\widetilde{\mathrm{I}}^{-}_{\widetilde{k}\rightarrow n} (\{{m}_i\}_{i \in \mathcal{I}_{\widetilde{k}}}) \nonumber\\
&\hspace{-0.5cm}\leftarrow
\vec{f}^{-}_{\widetilde{k}} ( (\{{m}_i\}_{i \in \mathcal{I}_{\widetilde{k}}}))
+\sum_{i \in \mathcal{I}_{\widetilde{k}} \setminus \{n\}}
{\mathrm{I}}_{i\rightarrow \widetilde{k}}({m}_{i}), \quad \text{if } \mathcal{D}(\widetilde{k},n), \\
&\hspace{-6.5cm}\widetilde{\mathrm{I}}^{+}_{\widetilde{k}\rightarrow n} (\{\widetilde{m}_i\}_{i \in \mathcal{I}_{\widetilde{k}}})\nonumber \\
&\hspace{-0.5cm}\leftarrow
\vec{f}^{+}_{\widetilde{k}} ( (\{\widetilde{m}_i\}_{i \in \mathcal{I}_{\widetilde{k}}}))
+\sum_{i \in \mathcal{I}_{\widetilde{k}} \setminus \{n\}}
\bar{\mathrm{I}}_{i\rightarrow \widetilde{k}}(\widetilde{m}_{i}), \quad \text{otherwise},
\end{eqnarray}
where $\bar{\mathrm{I}}_{n\rightarrow \widetilde{k}}(\widetilde{m}_{n})$ denotes the projected information from VN $n$ to FN $\widetilde{k}$. To adapt previous works \cite{iter-Mary, code-project} for AmBC over $L$ samples, we define FN indicator function as follows
\begin{eqnarray}
\Lambda(\widetilde{k},n)=\left\lfloor \sum_{i=1}^{\widetilde{k}} \widetilde{g}_{i,n}^{+}/L\right\rfloor+1.
\end{eqnarray}
Consequently, the information $\bar{\mathrm{I}}_{n\rightarrow \widetilde{k}}(\widetilde{m}_{n})$ is summarized in Table \ref{table:proj_ex}.
\begin{table*}
\small
\centering
\caption{Table of Information Projections and Expansions.}
\label{table:proj_ex}
\renewcommand{\arraystretch}{1.2} \setlength{\tabcolsep}{12pt}
\begin{tabular}{ccl}
\hline
{$M$} & {$\Lambda(\widetilde{k},n)$} & {\bf Projected Information from VN $n$ to FN $\widetilde{k}$} \\
\hline
\multirow{2}{*}{4} & $1$ & $\bar{\mathrm{I}}_{n \rightarrow \widetilde{k}} (1)=\mathrm{I}_{n \rightarrow \widetilde{k}} (1), \quad \bar{\mathrm{I}}_{n \rightarrow \widetilde{k}} (2)=\overset{\star}{\max}\left (\mathrm{I}_{n \rightarrow \widetilde{k}} (2),\mathrm{I}_{n \rightarrow \widetilde{k}} (3) \right),\quad \bar{\mathrm{I}}_{n \rightarrow \widetilde{k}} (3)=\mathrm{I}_{n \rightarrow \widetilde{k}} (4)$ \smallskip\\
{} & 2 & $\bar{\mathrm{I}}_{n \rightarrow \widetilde{k}} (1)=\mathrm{I}_{n \rightarrow \widetilde{k}} (2), \quad \bar{\mathrm{I}}_{n \rightarrow \widetilde{k}} (2)=\overset{\star}{\max}\left (\mathrm{I}_{n \rightarrow \widetilde{k}} (1),\mathrm{I}_{n \rightarrow \widetilde{k}} (4) \right),\quad \bar{\mathrm{I}}_{n \rightarrow \widetilde{k}} (3)=\mathrm{I}_{n \rightarrow \widetilde{k}} (3)$ \smallskip \\
\hline
\multirow{4}{*}{8} & \multirow{2}{*}{1} & $\bar{\mathrm{I}}_{n \rightarrow \widetilde{k}} (1)=\overset{\star}{\max} \left(\mathrm{I}_{n \rightarrow \widetilde{k}} (1),\mathrm{I}_{n \rightarrow \widetilde{k}} (2),\mathrm{I}_{n \rightarrow \widetilde{k}} (4)\right),
\quad \bar{\mathrm{I}}_{n \rightarrow \widetilde{k}} (2)=\overset{\star}{\max}\left (\mathrm{I}_{n \rightarrow \widetilde{k}} (3),\mathrm{I}_{n \rightarrow \widetilde{k}} (5) \right),$ \\
{} & {} & $\bar{\mathrm{I}}_{n \rightarrow \widetilde{k}} (3)=\overset{\star}{\max} \left(\mathrm{I}_{n \rightarrow \widetilde{k}} (6),\mathrm{I}_{n \rightarrow \widetilde{k}} (7),\mathrm{I}_{n \rightarrow \widetilde{k}} (8)\right)$ \smallskip\\
{} & \multirow{2}{*}{2} & $\bar{\mathrm{I}}_{n \rightarrow \widetilde{k}} (1)=\overset{\star}{\max} \left(\mathrm{I}_{n \rightarrow \widetilde{k}} (1),\mathrm{I}_{n \rightarrow \widetilde{k}} (5),\mathrm{I}_{n \rightarrow \widetilde{k}} (6)\right),
\quad \bar{\mathrm{I}}_{n \rightarrow \widetilde{k}} (2)=\overset{\star}{\max}\left (\mathrm{I}_{n \rightarrow \widetilde{k}} (2),\mathrm{I}_{n \rightarrow \widetilde{k}} (8) \right),$ \\
{} & {} & $\bar{\mathrm{I}}_{n \rightarrow \widetilde{k}} (3)=\overset{\star}{\max} \left(\mathrm{I}_{n \rightarrow \widetilde{k}} (3),\mathrm{I}_{n \rightarrow \widetilde{k}} (4),\mathrm{I}_{n \rightarrow \widetilde{k}} (7)\right)$ \smallskip \\
\hline
{$M$} & {$\Lambda(\widetilde{k},n)$} & {\bf Expanded Information from FN $\widetilde{k}$ to VN $n$}\\
\hline
\multirow{2}{*}{4} & 1 & $\bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (1)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(1), \quad \bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (2)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(2), \quad \bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (3)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(2), \quad \bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (4)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(3)$ \smallskip\\
{} & 2 & $\bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (1)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(2), \quad \bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (2)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(1), \quad \bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (3)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(3), \quad \bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (4)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(2)$ \smallskip\\
\hline
\multirow{4}{*}{8} & \multirow{2}{*}{1} & $\bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (1)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(1), \quad \bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (2)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(1), \quad \bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (3)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(2), \quad \bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (4)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(1),$ \\
{} & {} & $\bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (5)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(2), \quad \bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (6)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(3), \quad \bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (7)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(3), \quad \bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (8)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(3)$ \smallskip \\
{} & \multirow{2}{*}{2} & $\bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (1)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(1), \quad \bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (2)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(2), \quad \bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (3)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(3), \quad \bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (4)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(3),$ \\
{} & {} & $\bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (5)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(1), \quad \bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (6)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(1), \quad \bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (7)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(3), \quad \bar{\mathrm{I}}_{\widetilde{k} \rightarrow n} (8)=\vec{\mathrm{I}}^{+}_{\widetilde{k} \rightarrow n}(2)$\smallskip\\
\hline
\end{tabular}
\end{table*}
%
Then, the updated message sent to neighboring VNs from FN $\widetilde{k}$ can be efficiently approximated as
\begin{eqnarray}
&\hspace{-7.2cm}\vec{\mathrm{I}}_{\widetilde{k}\rightarrow {n}}^{-}(m_{n}) \nonumber\\ &\leftarrow
\overset{\star}{\max}_{m_{i},i \in \mathcal{I}_{\widetilde{k}} \setminus \{n\}}\left(\widetilde{\mathrm{I}}^{-}_{\widetilde{k}\rightarrow n} (\{{m}_i\}_{i \in \mathcal{I}_{\widetilde{k}}})\right), \quad \text{if } \mathcal{D}(\widetilde{k},n), \label{eqn:Ifvm}\\
&\hspace{-7.2cm}\vec{\mathrm{I}}_{\widetilde{k}\rightarrow {n}}^{+}(\widetilde{m}_{n}) \nonumber \\
&\leftarrow
\overset{\star}{\max}_{\widetilde{m}_{i},i \in \mathcal{I}_{\widetilde{k}} \setminus \{n\}}\left(\widetilde{\mathrm{I}}^{+}_{\widetilde{k}\rightarrow n} (\{\widetilde{m}_i\}_{i \in \mathcal{I}_{\widetilde{k}}})\right), \quad \text{otherwise}.
\label{eqn:Ifvp}
\end{eqnarray}
Based on the information \eqref{eqn:Ifvm} and \eqref{eqn:Ifvp}, VN $n$ sends the updated message to FN $\widetilde{k}$. Assuming $K_1=2$ and denoting the indices of FNs connected to VN $n$ as $\mathcal{J}_n$, the updated message can be represented by
\begin{eqnarray}
\mathrm{I}_{{n}\rightarrow \widetilde{k}}(m_n)&\hspace{-0.1in}\leftarrow \hspace{-0.1in}&
\ln ap_{n}(m_n)
+\mathrm{I}_{\widetilde{k}\rightarrow {n}}(m_{n})
\nonumber \\
&\hspace{-0.1in}-\hspace{-0.1in}&\overset{\star}{\max}_{m_n, i \in \mathcal{J}_n \setminus \{\widetilde{k}\}}\left(\mathrm{I}_{i\rightarrow {n}}(m_{n})\right)
\label{eqn:Ivf},
\end{eqnarray}
where the information from FN $\widetilde{k}$ to VN $n$ can be expressed by
\begin{eqnarray}
\mathrm{I}_{\widetilde{k}\rightarrow {n}}(m_n)=
\begin{cases}
\vec{\mathrm{I}}_{\widetilde{k}\rightarrow {n}}^{-}(m_{n}), \quad \text{if } \mathcal{D}(\widetilde{k},n), \\
\bar{\mathrm{I}}_{\widetilde{k}\rightarrow {n}}(m_{n}), \quad \text{otherwise},
\end{cases}
\end{eqnarray}
where $\bar{\mathrm{I}}_{\widetilde{k}\rightarrow {n}}(m_{n})$ denotes the expanded information which is also summarized in Table \ref{table:proj_ex}.
After $N_I$ iterations, VN $n$ estimates logarithm of the probability of the code $m_n$ as
\begin{eqnarray}
Q_n (m_n) \leftarrow \ln ap_{n}(m_n) + \sum_{i \in \mathcal{J}_n} \mathrm{I}_{i\rightarrow {n}}(m_{n}).
\end{eqnarray}
Finally, the log-likelihood ratio (LLR) for $m^{\dagger}$-th most significant bit (MSB) where $m^{\dagger} \in \{1,\cdots,\log_2 M\}$ for tag $n$ can be calculated as \cite{LDS}
\begin{eqnarray}
\Theta_{n,m^{\dagger}} &\hspace{-0.1in}\leftarrow\hspace{-0.1in}& \ln \left( \frac{p(d^{\dagger}_{n,m^{\dagger}})=0}{p(d^{\dagger}_{n,m^{\dagger}})=1}\right)
\nonumber \\&\hspace{-0.1in}=\hspace{-0.1in}& \overset{\star}{\max}_{m':d^{\star}_{m^{\dagger}}(m')=0}\left( Q_n (m')\right)
\nonumber \\&\hspace{-0.1in}-\hspace{-0.1in}&\overset{\star}{\max}_{m':d^{\star}_{m^{\dagger}}(m')=1}\left( Q_n (m')\right).
\end{eqnarray}
Then, backscatter data can be detected as
\begin{eqnarray}
\hat{d}_{n,m^{\dagger}}^{\dagger}=
\begin{cases}
1,\quad &\text{if } \Theta_{n,m^{\dagger}} \leq 0\\
0,\quad &\text{if } \Theta_{n,m^{\dagger}} > 0
\end{cases}, \quad \text{for } a_n = 1.
\end{eqnarray}
Based on the iterative detector utilizing MAI and ISI, high-rate backscatter can be achievable by exploiting inherent characteristic of sparsity in ambient energy harvesting as well as DBC providing additional diversity gain.
\\{\bf Lemma 1}
The worst-case complexity of the iterative detector is given by
\begin{eqnarray}
\mathcal{O}\left(\left(1-\frac{\widetilde{L}}{L}\right){\widetilde{M}}^{K-1}+\frac{\widetilde{L}}{L}M^{2K-3}\right) \label{eqn:complexity}
\end{eqnarray}
\begin{IEEEproof}
Information from FN to VN can be decomposed into two cases. In the first case where ISI is not affected, $K(L-\widetilde{L})$ FNs are connected to $K-1$ VNs, respectively, and $M$-ary codeword information is passed to VNs, resulting $M^{K-1}$ combination of information. If $M$-ary codeword is projected to $\widetilde{M}$ symbols, the size of information can be reduced to $\widetilde{M}^{K-1}$ using information projection \cite{iter-Mary, code-project}. On the other hand, $K\widetilde{L}$ FNs affected by ISI are connected to additional $K-2$ VNs in dyadic factor graph, resulting $M^{2K-3}$ combination of information. Since RF tag relies on energy harvesting, the complexity of detector diminishes as density of active tags becomes sparser.
\end{IEEEproof}
As expressed in \eqref{eqn:complexity}, ISI increases complexity of the iterative detector for achieving additional diversity gain. However, the fraction of the FNs affected by the interferences is very small when $\widetilde{L} << L$ for sufficiently long period of backscatter symbols. Therefore, we can enjoy low-complexity detection even for non-orthogonal and high-order AmBC.
At last, the average sum throughput of the proposed SC-AmBC can be evaluated as
\begin{eqnarray}
C &\hspace{-0.1in}=\hspace{-0.1in}&\lambda R_b p \left (\hat{ \mathcal{I}} \big| \mathcal{E}\right ) p \left ( \mathcal{E} \right),
\end{eqnarray}
where $R_b=R_s/K\log_2 M$ denotes the bit rate of RF tags and the successful information decoding event and energy harvesting event are denoted by $\hat{\mathcal{I}}$ and $\mathcal{E}$, respectively. The probability of successful energy harvesting and information decoding can be calculated as
\begin{eqnarray}
p\left(\hat{\mathcal{I}} \big | \mathcal{E} \right) &\hspace{-0.1in}=\hspace{-0.1in}& \mathbb{E}_{a_n=1} \left [
\frac{\left\|\hat{\mathbf{d}}_{n}^{\dagger} \odot \mathbf{d}_n^{\dagger} \right \|^2}{\log_2 M}\right], \label{eqn:BER}\\
p\left(\mathcal{E} \right) &\hspace{-0.1in}=\hspace{-0.1in}& \mathbb{E} \left [a_n\right], \label{eqn:HP}
\end{eqnarray}
where $\odot$ denotes exclusive-NOR operator.
In conclusion, the comparison of the conventional TD-AmBC and the proposed SC-AmBC is summarized in Table~\ref{table:comparison}.
\begin{table*}
\small
\centering
\caption{Comparison of the conventional TD-AmBC and the proposed SC-AmBC.}
\renewcommand{\arraystretch}{1.2} \setlength{\tabcolsep}{24pt}
\begin{tabular}{l l c c}
\hline
{} & {} & {\textbf{TD-AmBC}} & {\textbf{SC-AmBC}}\\
\hline
{\bf Modulation} & Load impedances & $M$ & $2$ \\
{} & Constellation dimension & $2$ & $2K$ \\
{} & Constellation symbols & $M$ & $\widetilde{M}$ \\
\hline
{\bf Detection} & ISI exploitation & no & yes \\
{} & MAI exploitation & no & yes \\
{} & Complexity & low & medium \\
{} & BER & high & low \\
\hline
{\bf Networking} & Max. sum throughput & $R_b$ & $\lambda R_b$ \\
{} & Max. connectivity & $K$ & $\binom{K}{K_1}$ \\
{} & Duty cycle & $1/K$ & $K_1/K$ \\
{} & Latency & high & low \\
{} & Multiple access & orthogonal & non-orthogonal \\
\hline
\end{tabular}
\label{table:comparison}
\end{table*}
\section{Simulation Results}\label{sec:simulation}
In this section, we evaluate the performance of our proposed SC-AmBC along with the conventional TD-AmBC. The simulation scenario is set as a wireless local area network (WLAN) using 2.4 GHz industrial, scientific and medical (ISM) band \cite{802.11}. Applications for the AmBC would be smart home or IoT environments where massive numbers of RF devices are seamlessly integrated into existing wireless infrastructure.
For AmBC, AP transmits IEEE 802.11g Wi-Fi OFDM signals \cite{backfi} while tags harvest these ambient RF signals \cite{EH-eff} and upload data streams by switching load impedances \cite{qam}.
Then, RF tag in SC-AmBC modulates $\widetilde{M}$ load impedances with non-orthogonal sparse codeword, while that in TD-AmBC modulates $M$ impedances with TDMA \cite{backfi,ofdm2}.
Then, the AP in SC-AmBC receives the illuminated signals experiencing fully-correlated DBC \cite{dyadic1,dyadic2,dyadic3,dyadic4,dyadic5,dyadic6,ofdm2} and detects data streams from these signals via the proposed D-MPA. For simplicity, we assume a system model with single antenna at AP and no forward-error correction (FEC) codes but extensions of our model to such scenarios are straightforward manner.
The AmBC is numerically evaluated by Monte-Carlo simulations, since it is challenging to derive a closed-form theoretical performance in multi-dimensional constellation \cite{scma3} and in the DBC \cite{dyadic6} especially for low SNR.
System parameters are summarized in Table~\ref{table:general_parameters}.
\begin{table*}
\small
\centering
\caption{System Paramters.}
\label{table:general_parameters}
\renewcommand{\arraystretch}{1.2} \setlength{\tabcolsep}{18pt}
\begin{tabular}{l l l}
\hline
{\textbf{Parameters}} & {\textbf{Values}} & {\textbf{References}}\\
\hline
OFDM signaling & $\sigma_s^2 = 20~{\mathrm{dBm}} ~/~ \sigma_n^2 = -90~{\mathrm{dBm}} ~/~T_s = 50~\mathrm{ns} $ & IEEE 802.11g \cite{802.11}\\
\hline
Antenna property & $A_e = 0.0012 ~\mathrm{m}^2~/~ G_s = +3~\mathrm{dB}$ & Testbed results \cite{backfi}\\
\hline
Tag hardware & $\eta = 0.25 ~/~ \sigma_c^2 = 0.58 ~\mathrm{pJ/symbol}$ & Testbed results \cite{qam,EH-eff}\\
\hline
Sparse code & $K_1 = 2 ~/~ N_I = 3~\mathrm{iterations}$ & Theoretical results \cite{scma3} \\
\hline
Coverage, rate & $d = 5~\mathrm{m}$ ~/~$L=8~\mathrm{samples}$ & Simulation settings \cite{backfi} \\
\hline
\end{tabular}
\end{table*}
\subsection{Applicability to $M$-ary Modulation}
\begin{figure*}
\begin{center}
\parbox{7.5cm}
{\small \centering \includegraphics [width=7.5cm]{Mod_HP} \\
(a) Harvesting probability $p\left(\mathcal{E} \right)$}
\quad
\parbox{7.5cm}
{\small \centering \includegraphics [width=7.5cm]{Mod_BER}
(b) Conditional BER $1-p\left(\hat{\mathcal{I}} \big | \mathcal{E} \right)$}
\vspace{-0.0cm}
\caption{Effect of the modulation order $M$ ($L^{+} = 3$, $K = 4$, $\lambda=1.5$ for SC-AmBC).}
\label{fig:mod}
\end{center}
\end{figure*}
\begin{figure*}[h]
\begin{center}
\parbox{7.5cm}
{\small \centering \includegraphics [width=7.5cm]{MP_HP} \\
(a) Harvesting probability $p\left(\mathcal{E} \right)$}
\quad
\parbox{7.5cm}
{\small \centering \includegraphics [width=7.5cm]{MP_BER}
(b) Conditional BER $1-p\left(\hat{\mathcal{I}} \big | \mathcal{E} \right)$}
\vspace{-0.0cm}
\caption{Effect of the number of the multipaths $L^{+}$ ($M=4$, $K = 4$, $\lambda=1.5$ for SC-AmBC).}
\label{fig:DBC}
\end{center}
\end{figure*}
\begin{figure*}[h]
\begin{center}
\parbox{7.5cm}
{\small \centering \includegraphics [width=7.5cm]{OL_HP} \\
(a) Harvesting probability $p\left(\mathcal{E} \right)$}
\quad
\parbox{7.5cm}
{\small \centering \includegraphics [width=7.5cm]{OL_BER}
(b) Conditional BER $1-p\left(\hat{\mathcal{I}} \big | \mathcal{E} \right)$}
\vspace{-0.0cm}
\caption{Effect of the number of the time slots $K$ ($M=2$, $L^{+} = 3$).}
\label{fig:OL}
\end{center}
\end{figure*}
\begin{figure*}[h]
\begin{center}
\parbox{7.5cm}
{\small \centering \includegraphics [width=7.5cm]{TH_TD} \\
(a) Average sum throughput $C$ (TD-AmBC)}
\quad
\parbox{7.5cm}
{\small \centering \includegraphics [width=7.5cm]{TH_SC}
(b) Average sum throughput $C$ (SC-AmBC)}
\vspace{-0.0cm}
\caption{Effect of the reflection coefficient $\alpha$ ($L^{+} = 3$).}
\label{fig:TH}
\end{center}
\end{figure*}
Fig.~\ref{fig:mod} shows the effect of the modulation order $M$ on the harvesting probability and the conditional BER.
In conventional TD-AmBC modulator, the harvesting probability is irrelevant to the value of $M$, since all the $M$-PSK symbols are on the circle of radius $\sqrt{\alpha}$ as shown in Fig.~\ref{fig:signal}. (a) with $\beta=0.25\alpha, \forall M$ in \eqref{eqn:NEPS}.
Note that the reflection coefficient $\alpha$ in the horizontal axis is related to SNR per symbol rather than SNR per bit.
Consequently, as $M$ increases, conditional BER in TD-AmBC degrades as the SNR per bit is reduced by factor $\frac{1}{\log_2 M}$ \cite{wireless}.
On the other hand, in SC-AmBC modulator, the harvesting probability is affected by the modulation order $M$ and the backscatter power $\sigma_b^2$ in \eqref{eqn:backpower}, which can be analyzed by the value of $\beta$ and the threshold $\theta$ in \eqref{eqn:EH-constraint}.
In 2-ary modulation, the normalized energy per symbol is given by $\beta = 0.5\alpha$, which is larger than those in TD-AmBC modulations.
Larger value of $\beta$ increases the threshold value $\theta$ and eventually reduces the harvesting probability.
Similarly, we can anticipate the performance degradation in 8-ary modulation with $\beta = 0.375\alpha$.
However, we can observe no performance degradation especially for 4-ary modulation.
In the 4-ary SC-AmBC, the duty cycle $D$ is doubled by using spreading sequences with factor $K_1=2$ and the backscatter power $\sigma_b^2$ is reduced to half by using zero-constellation point represented in Fig.~\ref{fig:signal}. (b).
As a result, the non-orthogonal signaling with zero symbols (i.e., idle state) in SC-AmBC can compensate the harvesting probability without energy loss.
In terms of the BER performance in Fig.~\ref{fig:mod}. (b), all the SC-AmBC schemes outperform the TD-AmBC schemes.
Extension of signal space via sparse coding can enable 1.5X overloading as well as achieving significant diversity gains.
Consequently, it is noted that for large $\alpha$, 4-ary SC-AmBC provides even lower BER than 2-ary TD-AmBC with same $\beta$.
Although the average backscatter power is the same, SC-AmBC can utilize instantaneous channel fluctuations among different time slots, achieving modulation diversity \cite{mod-div}.
Therefore, the proposed SC-AmBC can support small-form RF devices with $M$-ary modulation.
\subsection{Applicability to DBC}
Fig.~\ref{fig:DBC} shows the effect of the DBC for harvesting probability and conditional BER.
The harvesting probability increases as the number of multipath in forward channel, $L^{+}$ increases.
The multipath diversity reinforces the incident power at tag antenna in \eqref{eqn:incident}, improving efficiency of ambient energy harvesting in both the TD-AmBC and the SC-AmBC.
However, multipath is harmful to BER performance in conventional TD-AmBC.
To resolve ISI caused by DBC, simply filtering ISI-affected samples in detector \cite{backfi, ofdm2} incurs significant loss in SNR as the number of multipath $L^{+}$ increases.
Moreover, this limits the operational range of backscatter applications such as inventory control in warehouse and grocery stores \cite{drone}.
Thus, we can observe the tradeoff between the harvesting probability and the conditional BER in the TD-AmBC where ISI is not exploited.
On the other hand, the proposed SC-AmBC can fully exploit the ISI based on D-MPA and D-CEA.
Without these algorithms, we can partially exploit the ISI for performance improvement.
As shown in Fig.~\ref{fig:DBC}. (b), traditional MPA for one-way channel \cite{LDS, iter-Mary, code-project} shows worse BER than those of TD-AmBC, especially in DBC.
However, if we add the weighted-sum of ISI for information calculation in \eqref{eqn:fk_ISI} using D-MPA, the conditional BER decreases as $L^{+}$ increases, which is in sharp contrast to the conventional TD-AmBC detectors and the baseline SC-AmBC with MPA.
Therefore, the proposed D-MPA with D-CEA provides multipath diversity for both the energy harvesting and the information decoding by fully exploiting ISI in DBC.
\subsection{Applicability to NOMA}
Fig.~\ref{fig:OL} shows the effect of the number of time slots $K$ on the harvesting probability and the conditional BER.
Due to high computational complexity for large $K$, we use $M=2$ for this scenario.
As the number of time slots $K$ increases, the duty cycle decreases as $D = 1/K$ for TD-AmBC and $D=2/K$ for SC-AmBC. Consequently, backscatter power is reduced to $\beta = \frac{\alpha}{K}$ for TD-AmBC and $\beta = \frac{2\alpha}{K}$ for SC-AmBC, providing sufficient power to rectifier circuits in RF tags.
Remarkably, increasing $K$ reduces the performance gap between the TD-AmBC and the SC-AmBC.
In this manner, increasing $K$ improves the harvesting probability by reducing the duty cycle $D$ in Fig.~\ref{fig:OL}. (a).
On the other hand, increasing $K$ degrades the conditional BER as shown in Fig.~\ref{fig:OL}. (b).
Since the backscatter power $\beta$ is reduced in both the TD-AmBC and the SC-AmBC in \eqref{eqn:NEPS}, the BER performance is degraded accordingly.
It is noteworthy that the TD-AmBC detectors have smaller backscatter power (i.e., $\beta = \frac{\alpha}{K}$) than those of the SC-AmBC (i.e., $\beta = \frac{2\alpha}{K}$), exhibiting the worse BER in Fig.~\ref{fig:OL}. (b).
More importantly, the sparsity, caused by small duty cycle is not utilized in those TD-AmBC detectors.
Those detectors can only support OMA with limited number of RF tags (i.e., $\lambda = 1$) and thus are not suitable for massive connectivity.
The proposed SC-AmBC can efficiently utilize the sparsity for reliable AmBC detection as well as for overloading multiple RF tags (i.e., $\lambda > 1$), supporting NOMA. Although MAI may degrade the BER performance, low-complexity MPA achieves near-optimal detection performance by the virtue of the sparsity \cite{scma1},
As a result, the proposed SC-AmBC can enable massive and reliable connectivity over the conventional TD-AmBC.
\subsection{Applicability to Energy Harvesting}
Fig.~\ref{fig:TH} shows the effect of the reflection coefficient $\alpha$ for the average sum throughput.
The magnitude of reflection coefficient $\alpha$ balances the incident power arrived at tag antenna in \eqref{eqn:incident} between energy harvesting and information decoding. The energy-information tradeoff in AmBC determines the average sum throughput in various modulators and detectors \cite{ABCN-kaibin}.
However, in the conventional TD-AmBC, the throughput is only marginally improved as it does not exploit MAI and ISI by assuming orthogonality for the modulation and detection, as shown in Fig.~\ref{fig:TH}. (a).
The proposed SC-AmBC, on the other hand, is designed to exploit these interferences in non-orthogonal manner.
Fig.~\ref{fig:TH}. (b) demonstrates the benefits of sparse coding in the throughput.
Compared to $5.15$ Mbps throughput in TD-AmBC, the maximum throughput of SC-AmBC is $8.26$ Mbps, achieving additional 60\% gain.
In addition, the energy-information tradeoff can be flexibly adapted by controlling $K$, $M$, as well as $\alpha$ in SC-AmBC.
In conclusion, the proposed SC-AmBC is an energy-efficient solution for passive RF devices based on energy harvesting.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{Conclusions}\label{sec:conclusion}
In this paper, we have proposed AmBC modulator and detector based on sparse code to address low connectivity and channel fading issues inherent in AmBC.
The sparse code was employed at RF tags relying on ambient energy harvesting and supports non-orthogonal concurrent backscatter communications providing massive connectivity to Wi-Fi AP.
The sparse code using SC-BMA can be implemented to small-form factor RF devices and enables NOMA with $M$-ary modulation, achieving modulation diversity by extending the dimension of signal space.
For signal detection, the iterative D-MPA and the channel estimation based on D-CEA were proposed to overcome the DBC.
By exploiting MAI and ISI in modulation and detection, energy efficiency of AmBC can be enhanced, realizing massive connectivity for passive RF devices relying on energy harvesting.
Therefore, SC-AmBC can be applied to batteryless IoT including biosensors, wearables and smart homes where massive number of these RF devices are seamlessly integrated into existing wireless infrastructure.
% use section* for acknowledgment
\section*{Acknowledgment}
This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT) (NRF-2014R1A5A1011478).
\begin{thebibliography}{1}
% [1] Passive Wi-Fi
\bibitem{passive}
B. Kellogg, V. Talla, J.R. Smith and S. Gollakota,
"Passive Wi-Fi: Bringing low power to Wi-Fi transmissions,"
{\em in Proc. of the 13th Usenix Symposium on Networked Systems Design and Implementation (NSDI'16)},
Santa Clara, CA, pp. 151-164, Mar. 2016.
% [2] BackFi (SIGCOMM'15)
\bibitem{backfi}
D. Bharadia, K.R. Joshi, M. Kotaru, and S. Katti,
"BackFi: High throughput WiFi backscatter,"
{\em in Proc. of the ACM SIGCOMM'15},
London, United Kingdom, pp. 283-296, Aug. 2015.
% [3] BLE
\bibitem{ble}
J.F. Ensworth and M.S. Reynolds,
"Every smart phone is a backscatter reader: Modulated backscatter compatibility with Bluetooth 4.0 low energy (BLE) devices,"
{\em in Proc. of the 2015 IEEE International Conference on RFID (IEEE RFID 2015)},
San Diego, CA, pp. 78-85, Apr. 2015.
% [4] Drone Relay
\bibitem{drone}
Y. Ma, N. Selby and F. Adib,
"Drone relays for battery-free networks,"
{\em in Proc. of the ACM SIGCOMM'17},
Los Angeles, CA, USA, pp. 335-347, Aug. 2017.
% [5] Ambient Backscatter (SIGCOMM'13)
\bibitem{ambB}
V. Liu, A. Parks, V. Talla, S. Gollakota, D. Wetherall and J.R. Smith,
"Ambient backscatter: Wireless communication out of thin air,"
{\em in Proc. of the ACM SIGCOMM'13},
Hong Kong, China, pp. 39-50, Aug. 2013,
% [6] Wi-Fi Backscatter (SIGCOMM'14)
\bibitem{wifiB}
B. Kellogg, A. Parks, S. Gollakota, J.R. Smith and D. Wetherall,
"Wi-Fi backscatter: Internet connectivity for RF-powered devices,"
{\em in Proc. of the ACM SIGCOMM'14},
Chicago, IL, USA, pp. 607-618, Aug. 2014.
% [7] Turbocharging (SIGCOMM'14)
\bibitem{turbo}
A.N. Parks, A. Liu, S. Gollakota and J.R. Smith,
"Turbocharging ambient backscatter communication,"
\em{in Proc. of the ACM SIGCOMM'14},
Chicago, IL, USA, pp. 619-630, Aug. 2014.
% [8] 16-QAM
\bibitem{16qam}
S.J. Thomas and M.S. Reynolds,
"A 96 Mbit/sec, 15.5 pJ/bit 16-QAM modulator for UHF backscatter communication,"
\em{in Proc. of the 2012 IEEE International Conference on RFID (IEEE RFID 2012)},
Orlando, Florida, USA, pp. 185-190, Apr. 2012,
% [9] QAM
\bibitem{qam}
S. J. Thomas, E. Wheeler, J. Teizer and M.S. Reynolds,
"Quadrature amplitude modulated backscatter in passive and semipassive UHF RFID systems,"
{\em IEEE Trans. Microw. Theory Tech,} vol. 60, no. 4, pp. 1175-1182, Apr. 2012.
% [10] Retrodirective Beamforming
\bibitem{retro}
I. Krikidis,
"Retrodirective large antenna energy beamforming in backscatter multi-user networks,"
to appear in {\em IEEE Wireless Commun. Letters}, Feb. 2018.
% [11] Backscatter Relay
\bibitem{RelayB}
S. Gong, X. Huang, J. Xu, W. Liu, P. Wang, and D. Niyato,
"Backscatter relay communications powered by wireless energy beamforming,"
to appear in {\em IEEE Trans. Commun.}, Feb. 2018.
% [12] Dyadic 5 (2013)
\bibitem{dyadic5}
C. Boyer and S. Roy,
"Space time coding for backscatter RFID,"
{\em IEEE Trans. Wireless Commun.}, vol. 12, no. 5, pp. 2272-2280, May. 2013.
% [13] Dyadic 6 (2016)
\bibitem{dyadic6}
C. He, Z.J. Wang, C. Miao and V.C.M. Leung,
"Block-level unitary query: Enabling orthogonal-link space-time code with query diversity for MIMO backscatter RFID,"
{\em IEEE Trans. Wireless Commun.}, vol. 15, no. 3, pp. 1937-1949, Mar. 2016.
% [14] LoRa Backscatter
\bibitem{LoRaB}
V. Talla, M. Hessar, B. Kellogg, A. Najafi, J.R. Smith and S. Gollakota,
"LoRa backscatter: Enabling the vision of ubiquitous connectivity,"
{\em in Proc. of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies}, vol. 1, no. 3, pp. 105:1-105:24, Sep. 2017.
% [15] Rateless Code Backscatter (SIGCOMM'12)
\bibitem{rateless}
J. Wang, H. Hassanieh, D. Katabi and P. Indyk,
"Efficient and reliable low-power backscatter networks,"
{\em in Proc. of the ACM SIGCOMM'12},
Helsinki, Finland, pp. 61-72, Aug. 2012.
% [16] Inference
\bibitem{inference}
G. Zhu, S.-W. Ko and K. Huang,
"Inference from randomized transmissions by many backscatter sensors,"
{\em IEEE Trans. Wireless Commun.}, vol. 17, no. 5, pp. 3111-3127, May. 2018.
% [17] Time Hopping
\bibitem{TH}
W. Liu, K. Huang, X. Zhou and S. Durrani,
"Full-duplex backscatter interference networks based on time-hopping spread spectrum,"
{\em IEEE Trans. Wireless Commun.}, vol. 16, no. 7, pp. 4361-4377, Jul. 2017.
% [18] Dyadic 1 (2007)
\bibitem{dyadic1}
J.D. Griffin and G.D. Durgin,
"Link envelope correlation in the backscatter channel,"
{\em IEEE Commun. Letters}, vol. 11, no. 9, pp. 735-737, Sep. 2007.
% [19] Dyadic 2 (2008)
\bibitem{dyadic2}
J.D. Griffin and G.D. Durgin,
"Gains for RF tags using multiple antennas,"
{\em IEEE Trans. Antennas Propag.}, vol. 56, no. 2, pp. 563-570, Feb. 2008.
% [20] Dyadic 3 (2011)
\bibitem{dyadic3}
H.-C. Liu, W.-C. Lin, M.-Y. Lin and M.-H. Hsu,
"Passive UHF RFID tag with backscatter diversity,"
{\em IEEE Antennas Wireless Propag. Lett.}, vol. 10, pp. 415-418, May. 2011.
% [21] Dyadic 4 (2012)
\bibitem{dyadic4}
C. He, X. Chen, Z.J. Wang and W. Su,
"On the performance of MIMO RFID backscattering channels,"
{\em EURASIP J. on Wireless Commun. Netw.}, vol. 11, Nov. 2012.
% [22] Wireless Commun. (Goldsmith)
\bibitem{wireless}
A. Goldsmith,
{\em Wireless Communications}; Cambridge University Press: New York, NY, USA, 2005.
% [23] OFDM-B 2
\bibitem{ofdm2}
D. Darsena, G. Gelli and F. Verde,
"Modeling and performance analysis of wireless networks with ambient backscatter devices,"
{\em IEEE Trans. Commun.}, vol. 65, no. 4, pp. 1797-1814, Jan. 2017.
% [24] Cooperative Receiver (Y.C. Liang)
\bibitem{cooperative}
G. Yang, Y.-C. Liang and Q. Zhang,
"Cooperative receiver for ambient backscatter communications with multiple antennas,"
{\em in Proc. of the 2017 IEEE International Conference on Communications (ICC)},
Paris, France, May. 2017.
% [25] OFDM-B 1
\bibitem{ofdm1}
G. Yang and Y.-C. Liang,
"Backscatter communications over ambient OFDM signals: Transceiver design and performance analysis,"
{\em in Proc. of the 2016 IEEE Global Communications Conference (GLOBECOM)},
Washington, DC, USA, Dec. 2016.
% [26] Ternary MCS (Gongpu Wang)
\bibitem{codingB}
Y. Liu, G. Wang, Z. Dou and Z. Zhong,
"Coding and detection schemes for ambient backscatter communication systems,"
{\em IEEE Access}, vol. 5, pp. 4947-4953, Mar. 2017.
% [27] AmbB, TDMA (Xiao Lu)
\bibitem{ambB-TD}
X. Lu, D. Niyato, H. Jiang, D.I. Kim, Y. Xiao and Z. Han,
"Ambient backscatter assisted wireless powered communications,"
{\em IEEE Wireless Commun.}, vol. 25, no. 2, pp. 170-177, Jan. 2018.
% [28] SCMA1
\bibitem{scma1}
H. Nikopour and H. Baligh,
"Sparse code multiple access,"
{\em in Proc. of the 2013 IEEE 24th International Symposium on Personal Indoor and Mobile Radio Communications (PIMRC Workshops)},
London, United Kingdom, pp. 332-336, Sep. 2013.
% [29] SCMA2
\bibitem{scma2}
M. Taherzadeh, H. Nikopour, A. Bayesteh and H. Baligh,
"SCMA codebook design,"
{\em in Proc. of the 2014 IEEE 80th Vehicular Technology Conference (VTC Fall)},
Vancouver, BC, Canada, Sep. 2014.
% [30] Grant-free
\bibitem{grant-free}
K. Au, L. Zhang, H. Nikopour, E. Yi, A. Bayesteh, U. Vilaipornsawai, J. Ma and P. Zhu,
"Uplink contention based SCMA for 5G radio access,"
{\em in Proc. of the 2014 Globecom Workshops (GC Wkshps)},
Austin, Tx, USA, pp. 900-905, Dec. 2014.
% [31] M-ary modulation
\bibitem{iter-Mary}
Y. Wu, S. Zhang and Y. Chen,
"Iterative multiuser receiver in sparse code multiple access systems,"
{\em in Proc. of the 2015 IEEE International Conference on Communications (ICC)},
London, UK, pp. 2918-2923, Jun. 2015.
% [32] Projection Method
\bibitem{code-project}
A. Bayesteh, H. Nikopour, M. Taherzadeh, H. Baligh and J. Ma,
"Low complexity techniques for SCMA detection,"
{\em in Proc. of the 2015 Globecom Workshops (GC Wkshps)},
San Diego, CA, USA, Dec. 2015.
% [33] FBMC
\bibitem{FBMC}
R. Nissel, S. Schwarz and M. Rupp,
"Filter bank multicarrier modulation schemes for future mobile communications,"
{\em IEEE J. Sel. Areas Commun.}, vol. 35, no. 8, pp. 1768-1782, Aug. 2017.
% [34] ABCN Kaibin Huang
\bibitem{ABCN-kaibin}
K. Han and K. Huang,
"Wirelessly powered backscatter communication networks: Modeling, coverage and capacity,"
{\em IEEE Trans. Wireless Commun.}, vol. 16, no. 4, pp. 2548-2561, Apr. 2017.
% [35] Coded-QAM
\bibitem{codedqam}
C. Boyer and S. Roy,
"Coded QAM backscatter modulation for RFID,"
{\em IEEE Trans. Commun.}, vol. 60, no. 7, pp. 1925-1934, Jul. 2012.
% [36] SCMA3
\bibitem{scma3}
Y. Zhou, Q. Yu, W. Meng and C. Li,
"SCMA codebook design based on constellation rotation,"
{\em in Proc. of the 2017 IEEE International Conference on Communications (ICC)},
Paris, France, May. 2017.
% [37] Hybrid backscatter
\bibitem{hybridB}
V. Talla and J.R. Smith,
"Hybrid analog-digital backscatter: A new approach for battery-free sensing,"
{\em in Proc. of the 2013 IEEE International Conference on RFID (IEEE RFID 2013)},
Penang, Malaysia, pp. 74-81, Apr. 2013.
% [38] Full-Duplex, SIC
\bibitem{SIC}
D. Bharadia, E. McMilin and S. Katti,
"Full duplex radios,"
{\em in Proc. of the ACM SIGCOMM'13},
Hong Kong, China, pp. 375-386, Aug. 2013.
% [39] SCMA4
\bibitem{scma4}
J. Peng, W. Chen, B. Bai, X. Guo and C. Sun,
"Joint optimization of constellation with mapping matrix for SCMA codebook design,"
{\em IEEE Sig. Process. Letters}, vol. 24, no. 3, pp. 264-268, Mar. 2017.
% [40] Log-Domain MAP
\bibitem{log-MAP}
P. Robertson, E. Villebrun and P. Hoeher,
"A comparison of optimal and sub-optimal MAP decoding algorithms operating in the log domain,"
{\em in Proc. of the 1995 IEEE International Conference on Communications (ICC)},
Seattle, USA, pp. 1009-1013, Jun. 1995.
% [41] LDS CDMA
\bibitem{LDS}
R. Hoshyar, F.P. Wathan and R. Tafazolli,
"Novel low-density signature for synchronous CDMA systems over AWGN channel,"
{\em IEEE Trans. Sig. Process.}, vol. 56, no. 4, pp. 1616-1626, Apr. 2008.
% [42] Compressed Sensing
\bibitem{CS}
D.L. Donoho,
"Compressed sensing,"
{\em IEEE Trans. Inf. Theory}, vol. 52, no. 4, pp. 1289-1306, Apr. 2006.
% [43] Ambient Backscatter Analysis (Feifei Gao)
\bibitem{ambB-analysis}
G. Wang, F. Gao, R. Fan and C. Tellambura,
"Ambient backscatter communication systems: Detection and performance analysis,"
{\em IEEE Trans. Commun.}, vol. 64, no. 11, pp. 4836-4846, Nov. 2016.
% [44] Ambient Backscatter Noncoherent (Feifei Gao)
\bibitem{ambB-noncoherent}
J. Qian, F. Gao, G. Wang, S. Jin and H. Zhu,
"Noncoherent detections for ambient backscatter system,"
{\em IEEE Trans. Wireless Commun.}, vol. 16, no. 3, pp. 1412-1422, Mar. 2017.
% [45] Toeplitz Matrix
\bibitem{Toeplitz}
Z. Wang and G.B. Giannakis,
"Wireless multicarrier communications - where Fourier meets Shannon,"
{\em IEEE Sig. Process. Mag.}, vol. 17, no. 3, pp. 29-48, May. 2000.
% [46] EH eff
\bibitem{EH-eff}
Y.-H. Kim, H.-S. Ahn, C. Yoon, Y. Lim and S.-O. Lim
"An ambient RF energy harvesting and backscatter modulating tag system enabling zero-power wireless data communication,"
{\em in Proc. of the Seventh International Conference on the Internet of Things (IoT '17)},
Linz, Austria, Oct. 2017.
% [47] MQAM 1
\bibitem{mqam1}
J. Besnoff, M. Abbasi and D.S. Ricketts,
"High data-rate communication in near-field RFID and wireless power using higher order modulation,"
{\em IEEE Trans. Microw. Theory Techn.}, vol. 64, no. 2, pp. 401-413, Feb. 2016.
% [48] MQAM 2
\bibitem{mqam2}
M. Jordao, R. Correia and N.B. Carvalho,
"High order modulation backscatter systems characterization,"
{\em in Proc. of the 2018 IEEE Topical Conference on Wireless Sensors and Sensor Networks (WiSNet)},
Anaheim, CA, USA, pp. 44-46, Jan. 2018.
% [49] Modulation diversity
\bibitem{mod-div}
J. Boutros and E. Viterbo,
"Signal space diversity: A power- and bandwidth-efficient diversity technique for the Rayleigh fading channel,"
{\em IEEE Trans. Inf. Theory}, vol. 44, no. 4, pp. 1453-1467, Jul. 1998.
% [50] CS Survey
\bibitem{cs-survey}
J.W. Choi, B. Shim, Y. Ding, B. Rao and D.I. Kim,
"Compressed sensing for wireless communiciations: Useful tips and tricks,"
{\em IEEE Survey \& Tutorials}, vol. 19, no. 3, pp. 1527-1550, 2017.
% [51] WPCN channel reciprocity
\bibitem{wpcn}
G. Yang, C.K. Ho, R. Zhang and Y.L. Guan,
"Throughput optimization for massive MIMO systems powered by wireless energy transfer,"
{\em IEEE J. Sel. Areas Commun.}, vol. 33, no. 8, pp. 1640-1650, Aug. 2015.
% [52] 802.11g
\bibitem{802.11}
ISO/IEC standard for information technology - telecommunications and information exchange between systems - local and metropolitan area networks - specific requirements part 11: Wireless LAN medium
access control (MAC) and physical layer (PHY) specifications.
{\em ISO/IEC 8802-11 IEEE Std 802.11 Second edition 2005-08-01 ISO/IEC 8802 11:2005(E) IEEE Std 802.11i-2003 Edition}, pp. 1-721, 2005.
\end{thebibliography}
% biography section
%
% If you have an EPS/PDF photo (graphicx package needed) extra braces are
% needed around the contents of the optional argument to biography to prevent
% the LaTeX parser from getting confused when it sees the complicated
% \includegraphics command within an optional argument. (You could create
% your own custom macro containing the \includegraphics command to make things
% simpler here.)
%\begin{IEEEbiography}[{\includegraphics[width=1in,height=1.25in,clip,keepaspectratio]{mshell}}]{Michael Shell}
% or if you just want to reserve a space for a photo:
%
%\begin{IEEEbiography}{Michael Shell}
%Biography text here.
%\end{IEEEbiography}
%
%% if you will not have a photo at all:
%\begin{IEEEbiographynophoto}{John Doe}
%Biography text here.
%\end{IEEEbiographynophoto}
%
%% insert where needed to balance the two columns on the last page with
%% biographies
%%\newpage
%
%\begin{IEEEbiographynophoto}{Jane Doe}
%Biography text here.
%\end{IEEEbiographynophoto}
% You can push biographies down or up by placing
% a \vfill before or after them. The appropriate
% use of \vfill depends on what kind of text is
% on the last page and whether or not the columns
% are being equalized.
%\vfill
% Can be used to pull up biographies so that the bottom of the last one
% is flush with the other column.
%\enlargethispage{-5in}
% that's all folks
\end{document}