Scatter Radio Receivers for Extended Range Environmental Sensing WSNs
Extracted source text ยท arXiv 1703.07139
\documentclass[10pt,letterpaper]{IEEEtran}
\usepackage{graphicx}
\usepackage{amssymb}
\usepackage{amsmath}
\usepackage{amsfonts}
\usepackage{cite}
\usepackage{url}
\usepackage{enumerate}
\usepackage{epstopdf}
\usepackage{dsfont}
\usepackage{suffix}
\usepackage{steinmetz}
\usepackage{amsthm}
\usepackage{algpseudocode}
\usepackage{algorithm}
%\newtheorem{definition}{Definition}
\newtheorem{theorem}{Theorem}
\newtheorem{lemma}{Lemma}
%\newtheorem{corollary}{Corollary}
\newtheorem{proposition}{Proposition}
\newtheorem{conjecture}{Conjecture}
%\newtheorem{example}{Example}
%\newtheorem{remark}{Remark}
%\newcommand{\vectornorm}[1]{\left|\left|#1\right|\right|}
\newcommand{\HRule}{\rule{\linewidth}{0.5mm}}
%\newtheorem{definition}{Definition}
%\newtheorem{theor}{\textbf{Theorem}}
%\newtheorem{lem}{\textbf{Lemma}}
%
%\newtheorem{theorem1}{\textbf{Theorem}}
%\newtheorem{cor}{\textbf{Corollary}}
%\newtheorem{proposition}{Proposition}
%\newtheorem{example}{Example}
%\newtheorem{remark}{Remark}
%\newtheorem{property}{Property}
\makeatletter
\def\footnoterule{\relax%
\kern-5pt
\hbox to \columnwidth{\hfill\vrule width \columnwidth height 0.4pt\hfill}
\kern4.6pt}
\makeatother
\begin{document}
\title{ \vspace{-0.1 in} Scatter Radio Receivers for Extended Range
Environmental Sensing WSNs}
\vspace{-0.06 in}
\author{ \vspace{-0.08 in}
Panos~N.~Alevizos and %~\IEEEmembership{Student Member,~IEEE},
Aggelos~Bletsas % <-this % stops a space
\\
\IEEEauthorblockA{School of Electronic and Computer Engineering,
Technical University of Crete, Chania, Greece 73100\\
Email: palevizos$@$isc.tuc.gr, aggelos$@$telecom.tuc.gr }}
%
\maketitle
\vspace{-0.08 in}
\begin{abstract}
Backscatter communication, relying on the reflection principle,
constitutes a promising-enabling technology for
low-cost, large-scale, ubiquitous sensor networking.
This work makes an overview of the state-of-the-art coherent and noncoherent
scatter radio receivers that
account for the peculiar signal model
consisting of several microwave and communication parameters.
\end{abstract}
\vspace{-0.06 in}
\vspace{-0.06 in}
\section{Introduction}
%
The need of ubiquitous environmental sensing has lead to
the adoption of cost-effective, large-scale
wireless sensor networks (WSNs).
Such networks consists of several low-power and low-cost
devices that sense
environmental variables
(such as soil humidity, soil moisture, temperature) and gathering the sensed information at a central unit.
Existing commercial WSN equipment incorporates devices consisting of
complex active {radio frequency} (RF) components that
increase significantly the total monetary cost, as well
as the overall energy consumption per sensor node.
%
%
Scatter radio adopts the reflection principle \cite{Stockman:48},
which is achieved by generating a carrier wave that illuminates a set of RF tag/sensors. A RF tag terminates
its antenna load according to the data to be transmitted.
The incident signal is modulated and scattered back
towards a software-defined radio (SDR) reader for processing.
The above idea is depicted in Fig.~\ref{fig:sys_mod}.
This work proposes bistatic scatter radio technology with semi-passive
tags and frequency shift-keying (FSK) modulation, ideal for the power limited regime \cite{KiBlSa:14}.
The specific three-fold design mixture, not only reduces the energy cost per sensor tag,
but also does not sacrifice
in total communication range, achieving ranges on the order of a hundred of meters, as opposed to the conventional
passive radio frequency identification (RFID) Gen 2 architecture that offers a limited range on the order of very few meters.
A thorough overview of the state-of-the-art coherent and noncoherent
scatter radio FSK receivers is subsequently conducted. The receivers are capable to
achieve one hundred and fifty meters communication range.
\vspace{-0.06 in}
\section{Bistatic Scatter Radio Signal Model}
\label{sec:Signal Model}
The bistatic scatter radio architecture is employed, consisting of
a carrier emitter, a RF
tag, and a software-defined radio (SDR) reader
(Fig.~\ref{fig:sys_mod}).
Due to the relatively small bit-rate (on the order of few kilobits per second),
along with the fact that the carrier emitter-to-SDR reader and tag-to-SDR reader links are on the order of
a kilometer, i.e., small delay spread,
frequency non-selective (flat) fading channel is assumed, where the {baseband} complex channel response
for a duration of channel coherence time, $T_{\rm coh}$, is given by
$
h_{\rm k}(t) = h_{\rm k} = a_{\rm k} \mathsf{e}^{-\mathsf{j} \phi_{\rm k}} , ~ {\rm k}\in \{\rm CR,CT,TR\},
$
where $a_{\rm k}, \in \mathds{R}_+, ~{\rm k}\in \{\rm CR,CT,TR\},$ denote
the channel attenuation parameters of the corresponding links and
$\phi_{\rm k } \in [0,2\pi),~ {\rm k}\in \{\rm CR,CT,TR\},$ stand for
the respective phases due to signal propagation delay, all independent of each other.
Parameters $a_{\rm CR}, a_{ \rm CT}, a_{ \rm TR}$ are assumed
Rician distributed due to strong line-of-site signals,
commonly encountered in outdoors WSNs.
\begin{figure}[!t]
%\vspace{-0.12in}
\centering
%\vspace{-0.12 in}
\includegraphics[width=0.8\columnwidth]{bistatic_setup_2}
\caption{Bistatic architecture system model: carrier emitter is displaced from
SDR reader and RF tag modulates the incident RF signal from carrier emitter.}
\label{fig:sys_mod}
\end{figure}
Carrier emitter transmits a continuous sinusoid wave at carrier frequency $F_{\rm car}$
that illuminates the tag. Tag modulates the received signal at passband
using $50\%$ duty cycle square waveform pulse
of frequency $F_i$ and random initial phase $ \Phi_i \sim \mathcal{U}[0,2 \pi],~ i \in \mathds{B} \triangleq
\{0,1\}$ and backscatters it towards SDR reader.
The received complex baseband signal at the SDR reader is given
by the superposition of the carrier emitter sinusoid and the backscattered
tag signal through channels $h_{\rm CR}$ and $h_{\rm TR}$, respectively.
The received signal also suffers from band-limited noise
with power spectral density $N_0/2$ over SDR receiver bandwidth $ W_{\rm SDR}$.
SDR reader applies carrier frequency offset (CFO) estimation and compensation
using periodogram-based techniques, DC blocking, and synchronization.
For $|F_0-F_1|\gg \frac{1}{T}$ and $F_i \gg \frac{1}{T},$ the noise-free, CFO-free, DC-blocked, and perfect
synchronized baseband signal belongs to a four dimensional, time limited in $[0, T)$
signal space, whose orthonormal basis is denoted as
$\mathcal{B} \triangleq \left\{\frac{1}{\sqrt{T}} \mathsf{e}^{\pm \mathsf{j} 2\pi F_i t}\Pi_T(t)
\right\}_{i \in \mathds{B}}$, where
$\Pi_T(t)$ is the square waveform of duration $T$.
The optimal demodulator projects
received signal on basis $\mathcal{B}$ through correlators and the
discrete baseband signal over bit duration $T$ can be written as \cite{FaAlBl:15}
\begin{equation}
\vspace{-0.06 in}
\mathbf{r}
=
\left[ \begin{array}{ccc} r_0^+ \\ r_0^- \\
r_1^+ \\ r_1^- \end{array} \right] =
h \sqrt{\frac{\mathtt{E}}{2}} \left[ \begin{array}{ccc}\mathsf{e}^{+ \mathsf{j} \Phi_0 } \\ \mathsf{e}^{- \mathsf{j} \Phi_0 }
\\ \mathsf{e}^{+ \mathsf{j} \Phi_1 } \\\mathsf{e}^{- \mathsf{j} \Phi_1 } \end{array} \right]
\odot
\mathbf{s}_i
%\left[ \begin{array}{ccc} 1-i \\ 1-i
%\\ i \\ i \end{array} \right]
+
\left[ \begin{array}{ccc} n_0^+ \\ n_0^- \\
n_1^+ \\ n_1^- \end{array} \right],
\label{eq:discrete_equivalent}
\end{equation}
where $h = \alpha_{\rm CT}\alpha_{\rm TR}\mathsf{e}^{-\mathsf{j} \phi}$ incorporates the fading
coefficients $h_{\rm CT}h_{\rm TR} $ as well as
the phase difference of compound link
carrier emitter-to-tag-to-SDR reader.
$\mathtt{E}$ is the average energy per bit in baseband,
$\mathbf{s}_i = [(1-i)~ (1-i)~ i~ i]^{\top}$ is the four-dimensional
symbol corresponding to bit $i \in \mathds{B}$,
and
$
\mathbf{n} = [n_0^+ ~n_0^- ~n_1^+ ~n_1^-]^{\top} \sim \mathcal{CN}\left(\mathbf{0}_4,
\frac{N_0}{2} \mathbf{I}_4\right).
$
\vspace{-0.04 in}
\section{Bistatic Scatter Radio FSK Receivers}
\label{sec:Signal Model}
\subsection{Uncoded Reception}
\label{section:uncoded}
\subsubsection{Coherent Detector \cite{FaAlBl:15}}
The coherent receiver estimates the compound channel
$\mathbf{h}_{c} \triangleq h \sqrt{\frac{\mathtt{E}}{2}} \left[ \mathsf{e}^{+ \mathsf{j} \Phi_0 } ~
\mathsf{e}^{- \mathsf{j} \Phi_0 }~ \mathsf{e}^{+ \mathsf{j} \Phi_1 } ~ \mathsf{e}^{- \mathsf{j} \Phi_1 } \right]$
through the use of training symbols.
After obtaining the least-squares estimate of compound channel $\mathbf{h}_{c}$,
$\widehat{\mathbf{h}}_{c}$, the maximum-likelihood (ML)
detector is applied through the rule \cite{FaAlBl:15}
\begin{equation}
\widehat{i}_{\rm ML} = \arg \max_{i \in \mathds{B}} \Re \!\left\{
\left(\widehat{\mathbf{h}}_{c} \odot \mathbf{s}_i \right)^{H} \mathbf{r}
\right\}.
\end{equation}
\subsubsection{Noncoherent Detectors \cite{AlBlKar:17}}
The first symbol-by-symbol noncoherent detector
treats the parameter $h$ as deterministic and
parameters $\{\Phi_i\}_{i \in \mathds{B} }$ as random; it is called hybrid composite
hypothesis testing (HCHT) detector and is given by \cite{AlBl:15}
\begin{align}
\vspace{-0.06 in}
&~\arg \max \limits_{i \in \mathds{B}} \left\{ \underset{\Phi_i}{\mathbb{E}}\!\left[
\max \limits_{h \in \mathds{C}} \mathsf{ln}\! \left( \mathsf{f}(\mathbf{r} | i, h, \Phi_i) \right) \right] \right\}
\label{eq:hybrid_non_coh} \\
\Longleftrightarrow & ~ |r_0^+|^2 + |r_0^-|^2
\underset{i=1}{\overset{i=0}{\gtrless}}
|r_1^+|^2 + |r_1^-|^2 .
\label{eq:sqare_law_detector}
\vspace{-0.02 in}
\end{align}
The second symbol-by-symbol
detector is the generalized-likelihood ratio test (GLRT) detector, that treats all unknown parameters as deterministic and
is expressed as \cite{AlBlKar:17}
\begin{align}
\vspace{-0.06 in}
&~ \arg \max \limits_{i \in \mathds{B}} \left\{ \max\limits_{\Phi_i \in [0,2\pi]}
\max\limits_{h \in \mathds{C}} \mathsf{ln}\! \left( \mathsf{f}(\mathbf{r} | i, h, \Phi_i) \right) \right\}
\label{eq:GLRT_symbol_by_symbol} \\
\Longleftrightarrow & ~
|r_0^+| + |r_0^-|
\underset{i=1}{\overset{i=0}{\gtrless}}
|r_1^+| + |r_1^-| .
\label{eq:sum_abs_detector}
\vspace{-0.02 in}
\end{align}
For a bit sequence of $N$ bits satisfying $T_{\rm coh} \geq NT$, the received sequence a can be written as
$
\mathbf{r}_{1:N} = \left[ \mathbf{r}_1 ~ \mathbf{r}_2 ~ \ldots ~
\mathbf{r}_N \right] = h \left[ \mathbf{x}_{i_1}\! (\Phi_{i_1}) ~
\mathbf{x}_{i_2}\!(\Phi_{i_2})~\ldots ~
\mathbf{x}_{i_N}\!(\Phi_{i_N}) \right] + \left[ \mathbf{n}_1 ~\mathbf{n}_2 ~ \ldots ~
\mathbf{n}_N \right],$ with
$\mathbf{x}_{i_n} \! (\Phi_{i_n}) \triangleq \sqrt{\frac{\mathtt{E}}{2}} \left[ \mathsf{e}^{\mathsf{j}\Phi_0}~\mathsf{e}^{-\mathsf{j} \Phi_0}~
\mathsf{e}^{\mathsf{j}\Phi_1}~ \mathsf{e}^{-\mathsf{j}\Phi_1} \right]^{\top} \odot \mathbf{s}_{i_n},$
with $i_n \in \mathds{B}$,
$n=1,2,\ldots, N$. To reduce the sequence error rate, a noncoherent sequence detector
may be applied. For the above signal model, the GLRT sequence detector is expressed as
\begin{equation}
\arg \max\limits_{\mathbf{i} \in \mathds{B}^N} \left\{ \max\limits_{({\Phi}_0, \Phi_1) \in [0,2\pi]^2}
\max\limits_{h \in \mathds{C}} \mathsf{ln}\! \left( \mathsf{f}\! \left (\mathbf{r}_{1:N} | \mathbf{i} , h,
{\Phi}_0, \Phi_1 \right) \right) \right\}.
\end{equation}
Work in \cite{AlBlKar:17} partitions the space of
phases $({\Phi}_0, \Phi_1)$ in distinct $M^2$ points and solves the above problem
with complexity $\mathcal{O}(N\mathsf{log}(N))$ using the
same procedure with the work in \cite{AlFouKarBl:16}.
\subsection{Coded Reception}
\label{section:coded}
When channel coding is utilized, the transmitter
encodes a sequence of $K$ bits to a sequence
of $N$ coded bits, $\mathbf{c}=[c_1~c_2~\ldots,c_N]$ belonging to a
code $\mathcal{C} \subset \mathds{B}^N$.
\subsubsection{Coherent Decoder \cite{FaAlBl:15}}
After obtaining the least-squares estimate of compound channel $\mathbf{h}_{c}$,
$\widehat{\mathbf{h}}_{c}$, the ML decoder
can be expressed as \cite{FaAlBl:15}
\begin{equation}
\vspace{-0.06 in}
\widehat{\mathbf{c}}_{\rm ML} = \arg \max_{\mathbf{c} \in \mathcal{C}} \sum_{n=1}^N \Re \!\left\{
\left(\widehat{\mathbf{h}}_{c} \odot \mathbf{s}_{c_n} \right)^{H} \mathbf{r}_n
\right\}
\end{equation}
It can be shown that, if the above decoder is combined with interleaving
technique, it can achieve
diversity order $d_{\rm min}$, where $d_{\rm min}$
is the minimum distance of code $\mathcal{C}$ \cite{FaAlBl:15}.
\vspace{-0.02 in}
\subsubsection{Nonoherent Decoder \cite{AlBlKar:17}}
Suppose that we employ channel coding, noncoherent receiver, and the use of interleaving technique of
interleaving depth $D$
satisfying $D T \geq T_{\rm coh}$; accordingly with Eq.~\eqref{eq:hybrid_non_coh}, HCHT decoder
can be derived as \cite{AlBl:15}
\begin{equation}
\vspace{-0.05 in}
\widehat{\mathbf{c}} = \arg \max_{\mathbf{c} \in \mathcal{C}} \sum_{n=1}^N
\left(\left\|\mathbf{r}_n \odot \mathbf{s}_{1}\right\|^2-(\left\|\mathbf{r}_n \odot \mathbf{s}_{0}\right\|^2
\right) c_n .
\label{eq:comp_hyp_decoding rule}
\end{equation}
For the case of $D T< T_{\rm coh}$, the above decoder is suboptimal,
however, it is utilized for any value of $D$, due to its inherent simplicity.
\vspace{-0.07 in}
\section{Simulation Results}
\label{sec:simul_Results}
Simulations are conducted for bistatic scatter radio system model
assuming backscatter FSK transmissions with bit rate $T=1$msec, over quasi-static
Rician flat fading channel with coherence time $T_{\rm coh}=100$msec.
In all simulation results we have assumed perfect
synchronization, perfect CFO estimation/compensation,
and fixed energy budget per transmitted packet for all schemes.
%\begin{figure}[!t]
%\centering
% \includegraphics[width=\columnwidth]{NonCohUncoded_SimulTheory_v4}
%\caption{BER versus average received SNR of NC-CHT detection over Rayleigh fading, with
%simulation curve and analytical expression of Eq.~\eqref{eq:prob_error_Rayleigh} being perfectly matched.}
%\label{fig:NonCohUncoded_SimulTheory}
%\end{figure}
\begin{figure}[!t]
\centering
\vspace{-0.1 in}
\includegraphics[width=0.76\columnwidth]{NonCohSequence_vs_Coh_ImpacSequenceLength_v5}
% \includegraphics[width=0.85\columnwidth]{Figures/BER_coded_Coh_VS_NC_D_1_Impact_Fading_v6}
\caption{BER vs average received SNR for uncoded schemes. Due to fixed energy budget per transmitted packet and the fact that
coherent receiver employees extra $30$ bits for channel estimation (i.e., it
has less energy per bit), noncoherent GLRT sequence
detector outperforms coherent one.}
\label{fig:NonCohSequence_vs_Coh_ImpacSequenceLength}
\end{figure}
%\begin{figure}[!t]
%\vspace{-0.08 cm}
%\centering
%
%\caption{Impact of fading in BER performance of the noncoherent and coherent coded bistatic setup. Both setups
% utilize the cyclic $(31,11,11)$ BCH code and $D=1$.
% BER performance of both noncoherent and coherent receiver is affected significantly
%in highly random fading environments.}
%\label{fig:BER_coded_Coh_VS_NC_D_1_Impact_Fading}
%\end{figure}
Fig.~\ref{fig:NonCohSequence_vs_Coh_ImpacSequenceLength}
illustrates the bit error rate (BER) performance of all uncoded schemes studied in Section~\ref{section:uncoded}
as function of average received signal-to-noise ratio, $\frac{2\mathtt{E}}{N_0}$.
We observe that GLRT sequence
detector outperforms all schemes. It is also noted that coherent ML
detector offers 3dB better BER performance compared to noncoherent symbol-by-symbol schemes.
%Fig.~\ref{fig:NonCohSequence_vs_Coh_ImpacSequenceLength}-Right
%considers the use of $K=11$ and $N=31$ BCH channel code with interleaving depth $D=1$,
% and studies the impact of fading in BER of coded receivers presented in Section~\ref{section:coded}.
%It is remarked that as Rician parameters $\kappa_{\rm CT},\kappa_{\rm TR}$ increase, the BER gap between the two schemes is
%diminished.
\vspace{-0.05 in}
\bibliographystyle{IEEEtran}
\bibliography{IEEEabrv,BistaticBib_v2}
%figures
\end{document}
% Generated by IEEEtran.bst, version: 1.14 (2015/08/26)
\begin{thebibliography}{1}
\providecommand{\url}[1]{#1}
\csname url@samestyle\endcsname
\providecommand{\newblock}{\relax}
\providecommand{\bibinfo}[2]{#2}
\providecommand{\BIBentrySTDinterwordspacing}{\spaceskip=0pt\relax}
\providecommand{\BIBentryALTinterwordstretchfactor}{4}
\providecommand{\BIBentryALTinterwordspacing}{\spaceskip=\fontdimen2\font plus
\BIBentryALTinterwordstretchfactor\fontdimen3\font minus
\fontdimen4\font\relax}
\providecommand{\BIBforeignlanguage}[2]{{%
\expandafter\ifx\csname l@#1\endcsname\relax
\typeout{** WARNING: IEEEtran.bst: No hyphenation pattern has been}%
\typeout{** loaded for the language `#1'. Using the pattern for}%
\typeout{** the default language instead.}%
\else
\language=\csname l@#1\endcsname
\fi
#2}}
\providecommand{\BIBdecl}{\relax}
\BIBdecl
\bibitem{Stockman:48}
H.~Stockman, ``Communication by means of reflected power,'' \emph{Proc. IRE},
pp. 1196--1204, 1948.
\bibitem{KiBlSa:14}
J.~Kimionis, A.~Bletsas, and J.~N. Sahalos, ``Increased range bistatic scatter
radio,'' \emph{{IEEE} Trans. Commun.}, vol.~62, no.~3, pp. 1091--1104, Mar.
2014.
\bibitem{FaAlBl:15}
N.~Fasarakis-Hilliard, P.~N. Alevizos, and A.~Bletsas, ``Coherent detection and
channel coding for bistatic scatter radio sensor networking,'' \emph{{IEEE}
Trans. Commun.}, vol.~63, no.~5, pp. 1798--1810, May 2015.
\bibitem{AlBlKar:17}
P.~Alevizos, A.~Bletsas, and G.~N. Karystinos, ``Noncoherent short packet
detection and decoding for scatter radio sensor networking,'' \emph{{IEEE}
Trans. Commun.}, 2017, to appear.
\bibitem{AlBl:15}
P.~N. Alevizos and A.~Bletsas, ``Noncoherent composite hypothesis testing
receivers for extended range bistatic scatter radio {WSN}s,'' in \emph{Proc.
IEEE Int. Conf. on Commun. ({ICC})}, London, UK, Jun. 2015.
\bibitem{AlFouKarBl:16}
P.~N. Alevizos, Y.~Fountzoulas, G.~N. Karystinos, and A.~Bletsas,
``Log-linear-complexity {GLRT}-optimal noncoherent sequence detection for
orthogonal and {RFID}-oriented modulations,'' \emph{{IEEE} Trans. Commun.},
2016, to appear.
\end{thebibliography}