Iterative Decision Feedback Equalization Using Online Prediction
Extracted source text ยท arXiv 1810.03900
%% IterativeDecisionFeedbackEqualizationUsingOnlinePrediction.tex
%% V0.9a
%% 2016/01/25
%% by Serdar SAHIN
%% Template
%% 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/
\newif\ifdouble
%\doubletrue
\ifdouble
\documentclass[journal, 10pt, twoside]{IEEEtran}
\else
\documentclass[journal, 12pt, draftcls, onecolumn, twoside]{IEEEtran}
\fi
%
% If IEEEtran.cls has not been installed into the LaTeX system files,
% manually specify the path to it like:
% \documentclass[journal]{../sty/IEEEtran}
\usepackage[T1]{fontenc} % optional
% 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{{./diagrams/}{./images/}{./bio/}}
% and their extensions so you won't have to specify these with
% every instance of \includegraphics
\DeclareGraphicsExtensions{.pdf,.jpeg,.png,.eps}
\usepackage{epstopdf}
\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
% *** 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
\usepackage{algorithm}
\makeatletter
\newcommand\fs@norules{\def\@fs@cfont{\bfseries}\let\@fs@capt\floatc@ruled
\def\@fs@pre{}%
\def\@fs@post{}%
\def\@fs@mid{\kern3pt}%
\let\@fs@iftopcapt\iftrue}
\makeatother
%\floatstyle{norules}
\restylefloat{algorithm}
% *** 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. )
\newcommand{\specialcell}[2][c]{%
\begin{tabular}[#1]{@{}c@{}}#2\end{tabular}}
\usepackage{color}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{tikz}
\usepackage{hyperref}
\usepackage{tikz}
\usetikzlibrary{patterns}
\usetikzlibrary{backgrounds}
\usetikzlibrary{calc}
\usetikzlibrary{shapes}
\usetikzlibrary{fit}
\usetikzlibrary{chains}
\usetikzlibrary{arrows}
\usepackage{multirow}
\usepackage{tkz-graph}
\usetikzlibrary{intersections}
\usepackage{tkz-euclide}
\usetkzobj{all}
\usepackage{multirow}
\usepackage{tkz-graph}
%\usetikzlibrary{external}
%\tikzset{external/system call={latex \tikzexternalcheckshellescape -halt-on-error -interaction=batchmode -jobname "\image" "\texsource"&& dvips -o "\image".eps "\image".dvi}}
%\tikzexternalize[shell escape=-enable-write18]
\newcommand*{\Scale}[2][4]{\scalebox{#1}{$#2$}}%
\newcommand*{\Resize}[2]{\resizebox{#1}{!}{$#2$}}%
% \intersect{<p1>}{<p2>}{<q1>}{<q2>}
% draws the line p1--p2, showing a little semicircle
% where it intersects the line q1--q2.
\newcommand\intersect[4]{
\draw let \p{c} = (intersection of #1--#2 and #3--#4) in
(#1) -- ($(\p{c})!0.75mm!(#1)$)
to[bend right=90] ($(\p{c})!0.75mm!(#2)$) -- (#2)
}
% BLOCK DEFINITIONS
\tikzstyle{block} = [draw, rectangle, minimum height=2em, minimum width=0.5em]
%fill=blue!20
\tikzstyle{sum} = [draw, circle, node distance=1cm, fill=white]
%\tikzstyle{sum} = [draw, circle, node distance=1cm, fill=white, path picture={\draw (path picture bounding box.south east) -- (path picture bounding box.north west) (path picture bounding box.south west) -- (path picture bounding box.north east);}]
\tikzstyle{input} = [coordinate] \tikzstyle{output} = [coordinate]
\tikzstyle{pinstyle} = [pin edge={to-,thin,black}]
\usepackage[normalem]{ulem}
\newcounter{MYtempeqncnt}
% 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{\LARGE Iterative Decision Feedback Equalization Using Online Prediction}
\title{Iterative Decision Feedback Equalization Using Online Prediction}
%
%
% 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{Serdar~\c{S}ah\.{i}n*,~
Antonio~Maria~Cipriano,~
Charly~Poulliat,
and~Marie-Laure~Boucheret%
%Michael~Shell,~\IEEEmembership{Member,~IEEE,}
%John~Doe,~\IEEEmembership{Fellow,~OSA,}
%and~Jane~Doe,~\IEEEmembership{Life~Fellow,~IEEE}% <-this % stops a space
%\thanks{Manuscript received September XX, 2018. \emph{(Corresponding author: Serdar \c{S}ahin.)}}
\thanks{Manuscript sent to IEEE. \emph{(Corresponding author: Serdar \c{S}ahin.})}
\thanks{S. \c{S}ah\.{i}n is with Thales Communications and Security, 92230 Gennevilliers, France, and also with IRIT/INPT-ENSEEIHT, 31000 Toulouse, France (e-mail: serdar.sahin@thalesgroup.com).}%
\thanks{A. M. Cipriano is with Thales Communications and Security, 92230, Gennevilliers, France (e-mail: antonio.cipriano@thalesgroup.com).}%
\thanks{C. Poulliat and M.-L. Boucheret are with IRIT/INP Toulouse-ENSEEIHT, 31000, Toulouse, France (e-mails: charly.pouillat@enseeiht.fr, marie-laure.boucheret@enseeiht.fr).}%
\thanks{Digital Object Identifier XXXXXXXXXXXXX.}
}
%\thanks{M. Shell was with the Department
%of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta,
%GA, 30332 USA e-mail: (see http://www.michaelshell.org/contact.html).}% <-this % stops a space
%\thanks{J. Doe and J. Doe are with Anonymous University.}% <-this % stops a space
%\thanks{Manuscript received April 19, 2005; revised August 26, 2015.}}
% 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{Draft for IEEE - July 2019}%
{\c{S}ah\.{i}n \MakeLowercase{\textit{et al.}} - Iterative Decision Feedback Equalization Using Online Prediction}
% 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}
In this article, a new category of soft-input soft-output (SISO) minimum-mean square error (MMSE) finite-impulse response (FIR) decision feedback equalizers (DFEs) with iteration-wise static filters (i.e. iteration variant) is investigated.
It has been recently shown that SISO MMSE DFE with dynamic filters (i.e. time-varying) reaches very attractive operating points for high-data rate applications, when compared to alternative turbo-equalizers of the same category, thanks to sequential estimation of data symbols \cite{sahinCipriano_2018_DFEEP}.
However the dependence of filters on the feedback incurs high amount of latency and computational costs, hence SISO MMSE DFEs with static filters provide an attractive alternative for computational complexity-performance trade-off.
However, the latter category of receivers faces a fundamental design issue on the estimation of the decision feedback reliability for filter computation.
To address this issue, a novel approach to decision feedback reliability estimation through online prediction is proposed and applied for SISO FIR DFE with either a posteriori probability (APP) or expectation propagation (EP) based soft feedback.
This novel method for filter computation is shown to improve detection performance compared to previously known alternative methods, and finite-length and asymptotic analysis show that DFE with static filters still remains well-suited for high-spectral efficiency applications.
\end{abstract}
%\begin{keywords}
%Decision feedback equalizers, inter-symbol interference, expectation propagation, semi-analytical receiver abstraction, performance prediction, turbo equalization.
%\end{keywords}
% 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}
\label{sec:introduction}
{J}{oint} detection and decoding through iterative exchange of extrinsic information between a soft-input soft-output (SISO) detector and a SISO decoder can achieve near capacity performance with a well-designed coding scheme.
In particular, turbo equalization seeks to provide robust high-throughput links over strongly frequency-selective channels, whose frequency responses' incorporate spectral nulls.
However, unlike SISO finite-impulse response (FIR) minimum mean squared error (MMSE) linear equalizers (LE) \cite{tuchler_turbo_2011}, practical SISO MMSE FIR DFE structures have not been thoroughly investigated, and only gathered attention in recent years \cite{lou_soft_2014,tao_2016_low,sahinCipriano_2018_DFEEP}.
In this article, a novel filter computation approach for such structures is proposed, through a predictive estimation of the decision feedback reliability.
In the following, we only consider SISO MMSE FIR receiver structures.
A widespread and widely accepted nomenclature for categorizing such equalizers has not been established in the literature.
In our view, the work in \cite{jeong_2011_low_complexity_bidir} provides an accurate categorization based on the occurrence of SISO adaptive filter updates with prior information.
This nomenclature is attractive as it is directly related to the assumptions used for the derivation of the equalizer, and it also gives some insights on both the decoding performance and the computational complexity.
Time varying (TV) FIR filters are updated at each single symbol, by fully exploiting prior information, and they are well-suited for doubly-selective channels.
Iteration varying (IV) FIR filters are static and updated only at the beginning of each turbo iteration, by using the knowledge of overall quality of the feedback, thus reducing the involved computational costs.
The nature of feedback ``decisions'' (or rather ``estimates'') also impacts the DFE error-rate performance.
Hard decisions can be taken, as in conventional DFE \cite{belfiore_79_DFE}, but this results in a significant amount of error propagation and unpredictable behavior \cite{tuchler_turbo_2002}, unless complex heuristics are used \cite{jeong_turbo_2010}.
%by using symbol-wise likelihoods, as in conventional DFE \cite{belfiore_79_DFE}, but also based on extrinsic \cite{tuchler_turbo_2002} or posterior \cite{jeong_turbo_2010} probabilities.
Alternatively, a widespread category of DFE with soft feedback use a posteriori probability (APP) distributions, due to its relative simplicity and fair performance \cite{lopes_soft-feedback_2006,lou_soft_2014, tao_2016_low}.
Finally, soft feedback based on expectation propagation (EP) \cite{minka_2001_expectationpropagation} improves the performance of TV DFE \cite{sahinCipriano_2018_DFEEP}, in addition to being more predictable, due to relatively lowered correlations with the equalized estimates \cite{liWuTao_2019_perfAnalysisAndImprovemetForVAMPFDE}.
The design of optimum IV DFE receivers is however non-trivial. Indeed, static filters should depend on the decision feedback reliability, and the decision feedback naturally depends on the filters.
As a consequence, there does not exist closed-form expression of the optimal filter due to this non-linear ``chicken-and-egg'' inter-dependence.
IV DFE proposals in the literature use a variety of sub-optimal heuristics for filter computation \cite{lopes_soft-feedback_2006, lou_soft_2014, tao_2016_low}.
For DFE with hard decisions, the conventional approach is to assume a perfect feedback, causing error propagation and performance degradation in real operation \cite{tuchler_turbo_2002}.
Using soft APP decisions while still assuming perfect feedback partially mitigates error propagation, as soft symbols' magnitudes scale down with unreliability \cite{balakrishnan_mitigation_1999}.
The first reference to incorporate APP soft feedback reliability in filter computations is the receiver proposed in \cite{lopes_soft-feedback_2006} for the special case of binary-phase-shift-keying (BPSK) modulation.
The direct dependencies between soft symbols and log-likehood ratios (LLRs) for the BPSK constellation enable the use of a tractable density evolution on the APP LLR distribution, given a prior LLR distribution from the decoder.
This property is used by the BPSK receiver of \cite{lopes_soft-feedback_2006} to estimate the decision feedback reliability.
However, this scheme cannot be directly generalized to high-order constellations; hence \cite{lou_soft_2014} proposed a receiver that implements a LE at the first turbo iteration, and then it uses previous turbo iteration's demapper APP LLRs to estimate soft-symbol statistics for IV DFE.
Note that this approach is only possible with Gray mapped constellations.
More recently, \cite{tao_2016_low} proposed pre-equalization with LE over a few symbols, and then to use APP distribution of these symbols to estimate APP soft-feedback statistics for IV DFE.
Considering the previous developments, our contributions can be summarized as follows:
\begin{itemize}
\item An IV DFE with soft APP feedback based on online prediction is formulated.
The technique in \cite{lopes_soft-feedback_2006} is extended with models from the field of semi-analytic performance prediction \cite{visoz_semi-analytical_2010, sabbaghian_2009_analyticalapproach, yuan_2008_evolutionanalysis}, used both for physical layer abstraction or link layer adaptation.
\item A low complexity IV DFE with soft EP-based feedback is proposed, as an IV extension of \cite{sahinCipriano_2018_DFEEP}, by removing the bias from predicted APP estimates' reliability.
\item A novel approach for estimating soft decision feedback reliability is proposed, based on online binary and symbol-wise semi-analytic performance prediction.
\item The accuracy and the convergence of these prediction schemes are evaluated and the performance of IV equalizers are compared. To our knowledge, the accuracy of different IV DFE heuristics is not compared elsewhere, despite their direct impact on DFE error propagation.
\end{itemize}
The remainder of this paper is organized as follows. Section~\ref{sec:modelling} presents the system model and the general structure of proposed IV DFE receivers. The involved semi-analytic prediction scheme is developed and analyzed in Section~\ref{sec:abstraction}.
Finally, before concluding, the proposed equalizers with online prediction are numerically analyzed in Section~\ref{sec:perfos}, in the finite-length and the asymptotic regimes.
\begin{figure}[!t]
\centering
\includegraphics[width=0.48\textwidth]{fig_0_iter_recv_model.pdf}
%\scalebox{1}{\input{fig1_struct.tex}}
\caption{Iterative Detection and Decoding of BICM Signals.}
\label{fig_0_iter_recv_model}
\end{figure}
\subsection*{Notations}
Bold lowercase letters are used for vectors: let $\mathbf{u}$ be a $N \times 1$ vector, then $u_n, n=1,\dots,N$ are its entries, unless specified otherwise.
Capital bold letters denote matrices: for a $N \times M$ matrix $\mathbf{A}$, $[\mathbf{A}]_{n,:}$ and $[\mathbf{A}]_{:,m}$ respectively denote its $n^\text{th}$ row and $m^\text{th}$ column, and $a_{n,m}=[\mathbf{A}]_{n,m}$ is the entry $(n,m)$.
$\mathbf{I}_N$ is the $N\times N$ identity matrix, $\mathbf{0}_{N,M}$ and $\mathbf{1}_{N,M}$ are respectively all zeros and all ones $N\times M$ matrices. $\mathbf{e}_n$ is the $N\times 1$ indicator whose only non-zero entry is $e_n=1$.
Operator $\mathbf{Diag}(\mathbf{u})$ denotes the diagonal matrix whose diagonal is defined by $\mathbf{u}$.
$\mathbb{R}, \mathbb{C}$, and $\mathbb{F}_k$ are respectively real, complex and $k^\text{th}$ order Galois fields.
Let $x$ and $y$ be two random variables, $\mu_x=\mathbb{E}[x]$ is the expected value, $\sigma_x^2=\text{Var}[x]$ is the variance and $\sigma_{x,y}=\text{Cov}[x,y]$ is the covariance. The probability of the discrete random variable $x$ taking the value $\alpha$ is $\mathbb{P}[x=\alpha]$.
For random vectors $\mathbf{x}$ and $\mathbf{y}$, we define $\pmb{\mu}_\mathbf{x}=\mathbb{E}[\mathbf{x}]$ and covariance matrices $\mathbf{\Sigma}_{\mathbf{x},\mathbf{y}}=\mathbf{Cov}[\mathbf{x,y}]$ and $\mathbf{\Sigma}_{\mathbf{x}}=\mathbf{Cov}[\mathbf{x,x}]$.
$\mathcal{CN}(\mu_x,\sigma_x^2)$ denotes the circularly-symmetric complex Gaussian distribution of mean $\mu_x$ and variance $\sigma_x^2$.
\section{System Model}\label{sec:modelling}
\subsection{Single Carrier BICM Transmission}
Single carrier transmission using a bit-interleaved coded modulation (BICM) scheme is considered.
Let $\mathbf{b}\in\mathbb{F}_2^{K_b}$ be a $K_b$-bit information packet.
$\mathbf{b}$ is encoded and then interleaved into a codeword $\mathbf{d}\in \mathbb{F}_2^{K_d}$.
A memoryless modulator $\varphi$ then maps $\mathbf{d}$ to the symbol block $\mathbf{x}\in \mathcal{X}^K$, where $\mathcal{X}$ is the $M^\text{th}$ order complex constellation, with zero mean and average power $\sigma_x^2=1$, and with $Q=\log_2 M$.
The $Q$-word associated to $x_k$ is denoted $\mathbf{d}_k=[\mathbf{d}]_{Q(k-1)+1:Qk}$, and both $d_{k,q}$ and $\varphi_q^{-1}(x_{k})$ denote the value of the $q^\text{th}$ bit labelling the $x_{k}$, i.e. $d_{Q(k-1)+q}$, with $q=1,\dots,Q$.
We consider an equivalent baseband frequency selective channel with the impulse response $\mathbf{h}=[h_{L-1}, h_{L-2} \dots h_{0}]$, of delay spread $L$. Thus the received samples are
\begin{equation}
y_{k} = \sum_{l=0}^{L-1}{h_{l} x_{k-l}} + w_{k},
\end{equation}
for $k=1,\dots,K$, where the noise $w_k$ is modelled as an additive white Gaussian noise (AWGN), with $\mathcal{CN}(0, \sigma_w^2)$, i.e. a zero mean Gaussian process with variance $\sigma_w^2$.
The receiver is assumed to be ideally synchronized in time and frequency, and perfect channel state information is available.
We consider an iterative BICM receiver where a SISO channel decoder and a SISO symbol receiver exchange extrinsic information for iterative detection and decoding, as shown in Fig.~\ref{fig_0_iter_recv_model}.
A priori, extrinsic and a posteriori log likelihood ratios (LLRs) of coded bits $\mathbf{d}$ are respectively denoted $L_p(\cdot)$, $L_e(\cdot)$ and $L(\cdot)$, with respect to the SISO receiver.
The considered SISO symbol detector consists of a SISO channel equalizer and a symbol-wise SISO demapper module, as shown in Fig. \ref{fig_0_iter_recv_model}.
The latter consits in a soft-output maximum a posteriori (MAP) demapper, and a soft-mapping unit, an APP distribution estimator, and the eventual use of the so-called ``Gaussian division'' operation for computing extrinsic symbol feedback (see discussion below and \cite{sahinCipriano_2018_DFEEP}).
The SISO equalizer computes an estimate $x^e_k$ of $x_k$, affected by a residual noise of variance $v^e_{x,k}$, whereas the SISO demapper uses these estimates to compute $L_e(\mathbf{d})$, and to deliver soft feedback $x^d_k$ to the equalizer for additional interference cancellation, such that $v^d_{x,k}$ is the variance of residual interference and noise of the feedback (this will be discussed in more detail afterwards).
In short, soft mapper uses LLRs from the decoder to estimate a prior distribution on $x_k=\alpha$, $\forall \alpha\in\mathcal{X}$
\begin{equation}
\mathcal{P}_{k}(\alpha) \propto \textstyle\prod_{q=1}^{Q}e^{-\varphi^{-1}_q(\alpha)L_p(d_{k,q})}. \label{eq_priors}
\end{equation}
Soft demapper estimates a posteriori symbol distribution
\begin{equation}
\mathcal{D}_{k}(\alpha) \propto \exp{\left(-\frac{|\alpha-x^e_{k}|^2}{v_{x,k}^e}\right)}\mathcal{P}_{k}(\alpha),\, \forall \alpha\in\mathcal{X},
\label{eq_posteriors}
\end{equation}
which allows computing extrinsic LLRs towards the decoder
\begin{equation}
L_e(d_{k,j}) = \ln \frac{\sum_{\alpha\in\mathcal{X}_j^0}{\mathcal{D}_k(\alpha)}}{ \sum_{\alpha\in\mathcal{X}_j^1}{\mathcal{D}_k(\alpha)}} - L_p(d_{k,j}), \label{eq_demap_llr}
\end{equation}
% \begin{equation}
% L_e(d_{k,j}) = \textstyle\ln \sum_{\alpha\in\mathcal{X}_j^0}{\mathcal{D}_k(\alpha)} - \ln \sum_{\alpha\in\mathcal{X}_j^1}{\mathcal{D}_k(\alpha)} - L_p(d_{k,j}), \label{eq_demap_llr}
% \end{equation}
with $\mathcal{X}_j^b=\lbrace \alpha\in\mathcal{X}: \varphi^{-1}_j(x)=b\rbrace$ where $b\in\mathbb{F}_2$.
\subsection{On SISO FIR DFE Structures and Problem Statement}
FIR structures are modelled with windowed processes; applying a sliding window $[-N_p,N_d]$ on $y_k$, we define $\mathbf{y}_k=[y_{k-N_p},\dots,y_{k+N_d}]^T$.
$N_p$ and $N_d$ are respectively the number of pre-cursor and post-cursor samples, and we denote $N\triangleq N_p+N_d+1$, and $N_p'\triangleq N_p+L-1$.
Then, using the same window on $w_k$, and $[-N_p',N_d]$ on $x_k$, we have
\begin{equation}
\mathbf{y}_k = \mathbf{H}\mathbf{x}_k+\mathbf{w}_k, \label{eq_slidingwindow}
\end{equation}
with $\mathbf{H}$ being the $N\times N+L-1$ Toeplitz matrix generated by the static channel $\mathbf{h}$ with the first row being $\left[\mathbf{h}, \mathbf{0}_{1,N-1}\right]$.
Exact TV DFE receivers carry out interference cancellation with anti-causal estimates $\mathbf{\bar{x}}_k^{\mathbf{a}} \triangleq [\bar{x}^a_{k},\dots,\bar{x}^a_{k+N_d}]$ and causal estimates $\mathbf{\bar{x}}_k^{\mathbf{c}} \triangleq [\bar{x}^c_{k-N_p'},\dots,\bar{x}^c_{k-1}]$, with related variances $\mathbf{\bar{v}}_{\mathbf{x},k}^{\text{dfe}}\triangleq [ \bar{v}^c_{x,k-N_p'},\dots,\bar{v}^c_{x,k-1},\bar{v}^a_{x,k},\dots,\bar{v}^a_{x,k+N_d}]$.
Causal estimates are directly dependent on the values of $(x_k^e, v_{x,k}^e)$ from previous turbo iterations, and they depend on mapping constraints.
The equalized estimates are \cite{sahinCipriano_2018_DFEEP}
\begin{equation}
\begin{array}{l}
x^{e}_k = \bar{x}^{a}_k + \mathbf{f}_k^{}{}^H\mathbf{y}_k\\
{}\qquad - \mathbf{g}^{\mathbf{c}}_k{}^H\mathbf{\bar{x}}_k^{\mathbf{c}} - \mathbf{g}^{\mathbf{a}}_k{}^H\mathbf{\bar{x}}_k^{\mathbf{a}}\\
v^{e}_{x,k} = 1/\xi_k - \bar{v}^{a}_{x,k}
\end{array}
,\, \text{ }
\begin{cases}
\mathbf{f}_k \triangleq \mathbf{\Sigma}_k{}^{-1}\mathbf{h}_0/\xi_k, \\
\xi_k \triangleq \mathbf{h}_0^H\mathbf{\Sigma}_k^{-1}\mathbf{h}_0,
\end{cases}
\label{eq_dfetv_model}
\end{equation}
with
$\mathbf{\Sigma}_k \triangleq \sigma_w^2\mathbf{I}_N + \mathbf{H}\mathbf{Diag}(\mathbf{\bar{v}}^\text{dfe}_k)\mathbf{H}^H$, $\mathbf{g}^{\mathbf{c}}_k \triangleq [\mathbf{H}^H\mathbf{f}_k]_{1:N_p'}$, $\mathbf{g}^{\mathbf{a}}_k \triangleq [\mathbf{H}^H\mathbf{f}_k]_{N_p'+1:N_p'+N_d}$ and $\mathbf{h}_0 \triangleq [\mathbf{H}]_{:,N_p'+1}$.
In this paper, for numerical results the values of window parameters are taken as $N=3L+2$ and $N_d=2L$.
% figure DFE-IC invariant
\begin{figure*}[!t]
\centering
\includegraphics[width=0.55\textwidth]{fig_1_IVDFEIC_APP.pdf}
%\scalebox{1}{\input{fig1_struct.tex}}
\caption{Soft-Input Soft-Output IV DFE APP structure.}
\label{fig_1_IVDFEIC_APP}
\end{figure*}
IV DFE is obtained when $\mathbf{\bar{v}}^\text{dfe}_k$ is independent of $k$, with $\mathbf{\bar{v}}^\text{dfe}_k=\mathbf{\bar{v}}^\text{dfe}$, all filters being invariant, $\mathbf{f}$, $\mathbf{g}^{\mathbf{c}}$ and $\mathbf{g}^{\textbf{a}}$, as in \cite{tao_2016_low}:
\begin{equation}
\begin{array}{l}
x^{e}_k = \bar{x}^{a}_k + \mathbf{f}^{}{}^H\mathbf{y}_k\\
{}\qquad - \mathbf{g}^{\mathbf{c}}{}^H\mathbf{\bar{x}}_k^{\mathbf{c}} - \mathbf{g}^{\mathbf{a}}{}^H\mathbf{\bar{x}}_k^{\mathbf{a}}\\
v^{e}_x = 1/\xi - \bar{v}^{a}_x
\end{array}
,\, \text{ }
\begin{cases}
\mathbf{f} \triangleq \mathbf{\Sigma}{}^{-1}\mathbf{h}_0/\xi, \\
\xi \triangleq \mathbf{h}_0^H\mathbf{\Sigma}^{-1}\mathbf{h}_0.
\end{cases}
\label{eq_dfeiv_model}
\end{equation}
The variances of soft interference cancellation estimates are
\begin{equation}
\label{eq_iv_variance}
\mathbf{\bar{v}}^\text{dfe}_\mathbf{x} = [\bar{v}^c_x \mathbf{1}_{N_p',1}, \bar{v}^a_x \mathbf{1}_{N_d+1,1}],
\end{equation}
where $\bar{v}^a_x$ and $\bar{v}^c_x$ are respectively the overall reliability of anti-causal and causal estimates.
For interference cancellation, the set of anti-causal estimates are available before equalization, and an accurate estimate of their reliability is given by the least-squares estimation;
%\begin{equation}
$\bar{v}^a_x = K^{-1} \textstyle \sum_{k=0}^{K-1} \bar{v}^a_{x,k}$.
%\end{equation}
In most SISO DFE structures, the anti-causal estimates are the prior estimates given by the decoder\footnote{However, note that in the self-iterated SISO DFE of \cite{sahinCipriano_2018_DFEEP}, the anti-causal estimates are the causal estimates of previous iterations.}; ${x}^p_k \triangleq\mathbb{E}_{\mathcal{P}_k}[x_k], v^p_{x,k} \triangleq\text{Var}_{\mathcal{P}_k}[x_k]$.
% DEPLACER REPHRASER
As stated in the introduction, the core of the problem lies in the computation of $\bar{v}^c_x$.
A simple, but inaccurate solution to this is the ``perfect decision assumption''~: $\bar{x}^c_k$ are all assumed to be equal to $x_k$, yielding $\bar{v}^c_x=0$.
This approach is sufficient at high SNR operating points, but as shown in \cite{tuchler_turbo_2002,sahinCipriano_2018_DFEEP}, it degrades performance in moderately or severely selective channels.
Hence, with these notations, this paper's main objective is to evaluate novel prediction methods to compute $\bar{x}^c_k$ for optimizing the inter-symbol interference (ISI) mitigation performance of IV DFE.
In the following the cases of IV DFE based on APP-based soft feedback (see Fig.~\ref{fig_1_IVDFEIC_APP}), and EP-based soft feedback (see Fig.~\ref{fig_2_IVDFEIC_EP}) will be discussed.
%The case of a IV DFE based on EP feedback is shown in Fig. \ref{fig_dfe_prop}.
% figure DFE-IC invariant
\begin{figure*}[!t]
\centering
\includegraphics[width=0.65\textwidth]{fig_2_IVDFEIC_EP.pdf}
%\scalebox{1}{\input{fig1_struct.tex}}
\caption{Soft-Input Soft-Output IV DFE EP structure.}
\label{fig_2_IVDFEIC_EP}
\end{figure*}
\subsection{APP Soft Feedback Computation}\label{ssec:app_fb}
A common approach to compute soft feedback for DFE is to use APP estimates, as in IV structures of \cite{lopes_soft-feedback_2006,lou_soft_2014,tao_2016_low}. Derivation of a DFE APP for TV FIR is available in \cite{sahinCipriano_2018_DFEEP}.
These soft estimates are given by the mean and the variance of the posterior symbol distribution $\mathcal{D}_{k}$, given in Eq.~(\ref{eq_posteriors}). We denote
\begin{equation}
\label{eq_dem_stats}
\begin{aligned}
\mu^d_k&\triangleq\mathbb{E}_{\mathcal{D}_k}[x_k]&={}&\textstyle\sum_{\alpha\in\mathcal{X}}\alpha \mathcal{D}_k(\alpha),\\
\gamma^d_{x,k}&\triangleq\text{Var}_{\mathcal{D}_k}[x_k]&={}&\textstyle\sum_{\alpha\in\mathcal{X}}\vert\alpha\vert^2 \mathcal{D}_k(\alpha) - \vert \mu^d_k \vert^2.
\end{aligned}
\end{equation}
For the IV DFE APP filter computation, an invariant variance $\gamma^d_x$ is needed, as with the causal reliability $\bar{v}^c_x$.
Unlike the anti-causal reliability, $\gamma^d_x$ cannot be estimated using the causal estimates $(\bar{x}_k^c,\bar{v}^c_{x,k})=(\mu_k^d,\gamma_{x,k}^d)$, as causal estimates are only available once the filter is computed, and the equalization is being carried out.
Thus, a predictive estimation is required.
This receiver corresponds to the Fig.~\ref{fig_1_IVDFEIC_APP}, with $\bar{v}^c_x=\gamma^d_x$.
\subsection{Predictive EP-based Soft Feedback Computation}\label{ssec:ep_fb}
In \cite{sahinCipriano_2018_DFEEP}, an expectation propagation (EP) based soft feedback is used within a TV DFE which proved to bring several performance improvements.
Unlike APP estimates, EP-based estimates carry only the extrinsic information brought by the demapper, and prevents DFE from relying on its own bias, and improves the asymptotic predictability of the receiver.
For TV DFE, these estimates are obtained by the division of two Gaussian PDFs, which yields another Gaussian PDF with mean and variance given by
\begin{equation}
{x}^d_k = \frac{\mu^d_k v^e_x - x^e_k\gamma^d_{x,k}} {v^e_x-\gamma^d_{x,k}},\, \text{and},\, {v}^d_{x,k} = \frac{ v^e_x \gamma^d_{x,k}} {v^e_x-\gamma^d_{x,k}}.
\end{equation}
For the sake of simplicity, this operation is referred to as ``Gaussian division''.
For IV DFE EP structure, this feedback is not adapted, as the invariant filter is unable to adapt its coefficients to handle the strong variations of $v^d_{x,k}$, which depends on the instantaneous APP variance $\gamma_{x,k}^d$.
Hence we propose to use a feedback based on overall APP reliability, with
\begin{equation}
{x}^d_k = \frac{\mu^d_k v^e_x - x^e_k\gamma^d_x} {v^e_x-\gamma^d_x},\, \text{and},\, {v}^d_{x,k} = \frac{ v^e_x \gamma^d_x} {v^e_x-\gamma^d_x}, \label{eq_demap_ep_stats}
\end{equation}
where a predicted invariant APP variance $\gamma^d_x$ is used to generate the feedback.
Moreover EP-based estimates have an invariant variance, i.e. $v^d_x \triangleq v^d_{x,k}, \forall k$, as the causal reliability is directly related to the predicted APP variance.
The receiver based on this structure is illustrated in the Fig.~\ref{fig_2_IVDFEIC_EP}, with $\bar{v}^c_x=v^d_x$.
\section{Semi-Analytic Abstraction of SISO FIR DFE}\label{sec:abstraction}
In this section, a prediction model for the turbo DFE structure of Equation~(\ref{eq_dfeiv_model}) is exposed, without loss of generality, for the case where anti-causal estimates are given by decoder's extrinsic LLRs, i.e. $\bar{v}^a_x = v^p_x$.
Such models are usually used for handling physical layer link quality prediction that is necessary to enable link adaptation with low computational complexity. The originality is that, in our context, it will be exploited for online estimation of the reliability of causal estimates for SISO DFE filter computation.
\subsection{General Structure and Analytical Equalizer Model}
SISO DFE is modelled with two independent components; an analytical model for the equalizer, and a numerical model for the soft demapper.
Unlike asymptotic transfer models ($K \rightarrow + \infty$) used in extrinsic information transfer (EXIT) analysis \cite{brink_2000_designing}, finite-length transfer models are used for characterizing the demapper, as prior works on performance prediction noted their positive impact on accuracy \cite{sabbaghian_2009_analyticalapproach, visoz_semi-analytical_2010}.
\newpage
Following Eq.~(\ref{eq_dfeiv_model}), the IV DFE-IC output reliability $v^e_x$ is modelled by a function $\phi_\text{REC}$ as
\begin{equation}
\label{eq_rec_model}
\begin{split}
v^e_x &= \phi_\text{REC}(\sigma_w^2, \mathbf{h}, v^p_x , \bar{v}^c_x)\\
&\triangleq \left(\mathbf{h}_0^H \left[ \sigma_w^2\mathbf{I}_N + \mathbf{H}\mathbf{Diag}(\mathbf{\bar{v}}^\text{dfe}_\mathbf{x})\mathbf{H}^H \right]^{-1}\mathbf{h}_0 \right)^{-1} - v^p_x,
\end{split}
\end{equation}
where $\mathbf{\bar{v}}^\text{dfe}_\mathbf{x}$ is given by Eq.~(\ref{eq_iv_variance}), with $\bar{v}^a_x = v^p_x$.
This function is strictly increasing with $\bar{v}^c_x \in [0, \sigma_x^2]$.
As an analytical model is unavailable for characterizing the demapper, it is modelled with a look-up table (LUT) $\phi_\text{DEM}$, given by
\begin{equation}
\bar{v}^c_x = \phi_\text{DEM}(v^e_x, \cdot), \label{eq_dem_model}
\end{equation}
where $\bar{v}^c_x$ is the expected value of causal estimates' variance, taken over realizations of the channel noise, the equalizer outputs and the prior LLRs.
The second argument `$\cdot$' in Equation~(\ref{eq_dem_model}) models prior information, and the exact nature of the argument depends on the selected prediction approach.
Improvements proposed in the upcoming subsection concern this module.
Since equalizer and demapper iteratively exchange reliabilities, the two functions representing their model must be composed to yield a recursive equation on $\bar{v}^c_x$.
Hence by using $n=0,\dots,N_\text{pred}$ for indexing recursions, we have
\begin{equation}
\bar{v}^c_x[n+1] = \phi_\text{DEM}(\phi_\text{REC}(\sigma_w^2, \mathbf{h}, v^p_x, \bar{v}^c_x[n]), \cdot) \triangleq f_\text{pred}(\bar{v}^c_x[n]). \label{eq_fixedpoint}
\end{equation}
If $f_\text{pred}$ admits a unique fixed-point on $\bar{v}^c_x$, then
the desired predicted reliability estimate is this fixed-point. Moreover,
the optimality of IV DFE-IC strongly depends on $\bar{v}^c_x$ and hence on the accuracy of $\Phi_\text{DEM}$.
Fig.~\ref{fig_3_strct_pred} illustrates reliability prediction structures with the two semi-analytical models that will be introduced below.
\begin{figure}[!t]
\centering
\includegraphics[width=0.48\textwidth]{fig_3_strct_pred.pdf}
%\scalebox{1}{\input{fig2_model.tex}}
%\includegraphics[width=3.45in]{sahin2.eps}
\caption{Block diagrams for bit-wise (left) and symbol-wise (right) causal reliability prediction schemes.}
\label{fig_3_strct_pred}
\end{figure}
\subsection{Numerical Demapping Models for APP/EP}
Modelling the demapper with prior information is challenging due to its highly non-linear behavior, and due to strong simplifying assumptions.
The main focus of the numerical model will be the model of the posterior symbol distribution's variance $\gamma^d_x$, required for using the DFE with APP feedback, discussed in Subsection~\ref{ssec:app_fb}.
Besides, as the variance $v^d_x$ of the proposed predictive EP feedback, in Subsection~\ref{ssec:ep_fb}, is analytically linked to $\gamma^d_x$, numerical model of APP estimates is common for both types of feedback.
\subsubsection{Mutual Information based Prediction (Bit-wise)}
In the BPSK receiver of \cite{lopes_soft-feedback_2006}, a prediction scheme is considered, assuming input/output LLRs of the demapper to be consistent Gaussian, i.e. $L_{(\cdot)}(d_{k,q})\sim \mathcal{N}(\bar{d}_{k,q}\mu_{(\cdot)}, 2\mu_{(\cdot)})$, where $\bar{d}_{k,q}=1-2d_{k,q}$, and where $(\cdot)$ is $p, e$ or void, depending on concerned LLRs.
Using a semi-analytical density evolution, parameter $\mu_e$ of extrinsic LLRs is predicted using $\mu_p$.
The parameter $\mu_{(\cdot)}$ is bijectively linked to the average mutual information (MI) between LLRs and the associated coded bits, usable for binary prediction as in \cite{visoz_semi-analytical_2010}.
Hence using such formalism, the approach of \cite{lopes_soft-feedback_2006} can be extended to any constellation and mapping.
More specifically, the demapper behaviour is numerically integrated for each $\gamma_{x,k}^d$, $k=1,\dots,K$, under the assumption that prior LLRs are consistent Gaussian with the parameter $\mu_p$, and the assumption that the residual ISI and noise affecting the equalized symbols $x_k^e$ are Gaussian-distributed, i.e. $x_k^e \sim \mathcal{CN}(x_k, v^e_x)$.
Hence with these conditions, a LUT on $\mu_p$ and $v^e_x$ is built with
\begin{align}
\bar{v}^c_x = \phi_\text{DEM}(v^e_x, \mu_p) \triangleq \frac{1}{K} \sum_{k=1}^{K} \mathbb{E}_{\mathbf{L}_\mathbf{p}, x^e}[\bar{v}^c_{x,k}], \label{eq_feedback_expectation}
\end{align}
where $\bar{v}^c_x$ is the variance of residual error on APP/EP soft symbols and the priors' parameter $\mu_p$ is measured with a maximum-likelihood (ML) estimator (see Fig. \ref{fig_3_strct_pred}, left)
\begin{equation}
\mu_p \approx \sqrt{1+\sum_{k=1}^{K}\sum_{q=1}^{Q}\lvert L_p(d_{k,q})\rvert^2}-1.
\end{equation}
In the case of APP feedback, i.e. for $\bar{v}^c_{x,k}=\gamma_{x,k}^d$, the expectation in Equation~(\ref{eq_feedback_expectation}) becomes
\begin{align}
&\textstyle \mathbb{E}_{\mathbf{L}_\mathbf{p}, x^e}[\gamma_{x,k}^d] = \frac{1}{M}\sum_{x_k \in \mathcal{X}} \text{Var}\left[\mathcal{D}(x_k,x^e_k,\mathbf{L}_{\mathbf{p},k})\right] \nonumber\\
&\textstyle \quad \mathcal{CN}(x_k^e;x_k, v^e_x) \prod_{q=1}^Q \mathcal{N}(L_{p,k,q};\bar{\varphi}^{-1}_q(x_k)\mu_p,2\mu_p)
\end{align}
where the APP probability mass function of a dummy symbol $x\in\mathcal{X}$ is
\begin{equation}
\mathcal{D}(x,x^e,\mathbf{L}_{\mathbf{p}}) \triangleq \frac{1}{Z} \exp{\left(-\frac{|x-x^e|^2}{v^e} - \sum_{q=1}^{Q}\varphi^{-1}_q(x^e)L_{p,q} \right)}, % \prod_{q=1}^{Q}e^{-\varphi^{-1}_q(x^e)L_{p,q}}
\end{equation}
where $Z$ is the normalization constant such that $\sum_{x \in \mathcal{X}} \mathcal{D}(x,x^e,\mathbf{L}_{\mathbf{p}})=1$.
Moreover, considering the EP feedback's analytical expression in Equation~(\ref{eq_demap_ep_stats}), we have
\begin{equation}
\mathbb{E}_{\mathbf{L}_\mathbf{p}, x^e}[\gamma_{x,k}^d] = \left( \left( \frac{1}{K} \sum_{k=1}^{K}\mathbb{E}_{\mathbf{L}_\mathbf{p}, x^e}[\gamma_{x,k}^d] \right)^{-1} - \frac{1}{v^e_x}\right)^{-1}.
\end{equation}
The binary prediction scheme above appeared to yield too optimistic estimates in \cite{lopes_soft-feedback_2006}, and instead \emph{Lopes et al.} resorted to obtain $\mu_e$ and $\mu_p$ through BPSK channel estimators, which circumvents consistent Gaussian LLR approximation.
More specifically, this problem ensues from well known issues with regards to performance prediction of turbo iterative systems, for which the consistent Gaussian approximation of LLRs was shown to be only accurate at the zeroth turbo-iteration, and in the asymptotic limit.
Otherwise inaccurate estimates propagate across turbo iterations due to the internal non-linear dynamics of channel decoding \cite{fu_2005_stochasticanal}.
To overcome this prediction bias, prediction based on a two-parameter LLRs' model has been shown to be much more accurate \cite{ibing_2010_predictingconvergence}.
Such models consider $L_{(\cdot)}(d_{k,q})\sim \mathcal{N}(\bar{d}_{k,q}\mu_{(\cdot)}, \eta_{(\cdot)}\mu_{(\cdot)})$, where $\eta_{(\cdot)}$ is no longer 2.
The ML estimator used for measuring $\mu_p$ in the binary prediction is very sensitive to $\eta_{p}$, which is the reason why the binary prediction is not robust enough in practice.
\begin{figure}[t]
\centering
\includegraphics[width=0.48\textwidth]{fig_4_pred_quality.pdf}
\caption{Mean-square error on the prediction quality of the binary (blue, $\times$) and symbol-wise (red, +) schemes.}
\label{fig_4_pred_quality}
\end{figure}
\begin{figure}[t]
\centering
\includegraphics[width=0.45\textwidth]{fig_5_pred_fixed_pt.pdf}
\caption{Fixed-points of the symbol-wise $f_\text{pred}$ for SNR varying from 6 to 10 dB, for each value of prior reliability.}
\label{fig_5_pred_fixed_pt}
\end{figure}
\subsubsection{Prior Variance based Prediction (Symbol-wise)}
For our context, two-parameter models are too complex as they require expensive online parameter estimators to get both $\mu_p$ and $\eta_p$.
Hence, a single-parameter demapper model with reasonable estimation complexity has been preferred.
We searched for the parameter which is the least sensitive to the variations of prior LLRs' variance-to-mean ratio $\eta_{p}$.
Following a thorough and almost exhaustive study of the different alternative parameters for tracking evolution of $\bar{v}^c_x$, anti-causal variance $v^p_x$ has been found to be sensitive to the changes on $\eta_p$, very similarly to $\bar{v}^c_x$, with the advantage of $v^p_x$ being directly computable online using a simple least-squares estimation. Hence, we propose the following LUT
\begin{equation}
\bar{v}^c_x = \phi_\text{DEM}(v^e_x, v^p_x),\,\textstyle \text{ }
\begin{cases}
\bar{v}^c_x \triangleq K^{-1} \sum_{k=1}^{K} \mathbb{E}_{\mathbf{L}_\mathbf{p}, x^e}[\bar{v}^c_{x,k}], \\
v^p_x \triangleq K^{-1} \sum_{k=1}^{K} \mathbb{E}_{\mathbf{L}_\mathbf{p}}[v^p_{x,k}],
\end{cases}
\end{equation}
where both input $v^p_x$ and output $\bar{v}^c_x$ are numerically integrated using prior LLRs generated for a fixed value of $\eta_p$.
Indeed, as $\eta_p$ cannot be accurately measured online, the conventional consistent approximation \cite{visoz_semi-analytical_2010} is kept with $\eta_p=2$.
In the following, we will assess its impact on the prediction accuracy.
\subsubsection{Robustness of demapper prediction}
The sensitivity of the considered prediction schemes to variations in $\eta_{p}$ is evaluated. This aspect is important for characterizing the robustness of iterative receiver prediction schemes, as the hypothesis $\eta_{p}=2$, used for LUT generation, is only true at the initial turbo-iteration and then it varies \cite{fu_2005_stochasticanal}.
\begin{figure*}[t]
\centering
\includegraphics[width=0.695\textwidth]{fig_6_pred_conv_speed.pdf}
\caption{Evolution of APP covariance prediction $\gamma^d_{n+1} = f_\text{pred}(\gamma^d_{n})$. Different colors of plots are for SNR varying from -5 to 25 dB, with 5 dB steps (towards lower plots).}
\label{fig_6_pred_conv_speed}
\end{figure*}
An AWGN channel is simulated with blocks of 16-QAM symbols with $K=1024$, to emulate the output $x^e$ of the equalizer, for $v^e$ varying from $-15$ to $15$~dB, along with Gaussian-distributed prior LLRs generated with prior MI $I_A$ varying from 0 to 1 bit (and hence determining $\mu_p$), with $\eta_p$ varying from 1 to 3. The average mean squared error (MSE) between the predicted causal covariance and true causal covariance is measured, and plotted in Fig.~\ref{fig_4_pred_quality}.
The left side of the figure provides results for APP feedback, and the right side for EP-based feedback.
The binary approach is seen to be severely impacted by the changes in $\eta_p$, whereas the symbol-wise approach, although not perfect, remains more robust. Considerable differences are seen at low to medium SNR for high prior information, which suggests that symbol-wise schemes would have an advantage at the decoding threshold in asymptotic behaviour, i.e. when a high number of turbo-iterations are used. Oppositely, without any turbo-iteration, both schemes would perform identically.
\subsection{Convergence Analysis}
The convergence of the proposed iterative semi-analytical prediction schemes could be assessed formally through fixed-point analysis of Eq. (\ref{eq_fixedpoint}).
However, due to the untractable non-linear expression of $\Phi_\text{DEM}$, an analytic approach is not possible, and we resort to numerical evaluations.
In the following, the convergence of the symbol-wise prediction scheme is evaluated.
Numerical evaluations of the proposed $f_\text{pred}$ show that we can reasonably conjecture that this function is continuous on the interval $[0,+\infty[$, with a Lipschitz constant strictly less than one, for all $\sigma_w^2 \geq 0$ and $0 \leq v^p_x \leq \sigma_x^2$. This ensures that Eq. (\ref{eq_fixedpoint}) reaches a unique fixed-point $\bar{v}^c_x \in [0,+\infty[$ for any initial guess.
This conjecture has been checked for various channels $\mathbf{h}$, and Fig.~\ref{fig_5_pred_fixed_pt} plots $f_\text{pred}$ for the Proakis-C channel ($\mathbf{h}=[1,2,3,2,1]/\sqrt{19}$), using the symbol-wise demapper model for the Gray-mapped 16-QAM constellation.
The convergence speed of the prediction scheme is also evaluated numerically in order to determine optimal parameters of the algorithm.
Indeed, the fixed-point $\bar{v}^c_x=\bar{v}^c_x[\infty]$ is reached more or less quickly depending on if the initial value $\bar{v}^c_x[0]$ is close to $\bar{v}^c_x[\infty]$.
In particular, due to the near flat evolution of $f_\text{pred}$ for $\bar{v}^c_x$ close to $\sigma_x^2=1$, initializing with $\bar{v}^c_x[0]=1$ results in fast convergence at low SNRs, and high anti-causal covariance, but slow convergence otherwise.
Oppositely with $\bar{v}^c_x[0]=0$ faster convergence is achieved for high SNRs and low anti-causal covariance.
This behaviour is illustrated for Proakis-C 16-QAM APP covariance in Figure~\ref{fig_6_pred_conv_speed}.
We propose to use the heuristic $\bar{v}^c_x[0]= \min(1, \sigma_w)$, when $v^p_x>0.5$, where the standard deviation of the channel noise is experimentally shown to serve as a convergence accelerating heuristic, when the channel is normalized.
Otherwise using $\bar{v}^c_x[0]=0$ is preferable for faster convergence.
\begin{figure*}[t]
\centering
\includegraphics[width=\textwidth]{fig_7_uncoded_ber.pdf}
\caption{Uncoded bit-error rate (BER) performance of proposed predictive IV DFEs.}
\label{fig_ber_uncode}
\end{figure*}
\section{Numerical Results}\label{sec:perfos}
\subsection{Uncoded equalization behaviour}
In this paragraph, the uncoded finite-length behaviour of the proposed IV DFE with online prediction is evaluated.
Exact TV DFE counterparts are used as lower-bound references on bit-error rate (BER), to assess the prediction accuracy. Note IV FIR receivers might outperform TV FIRs in some cases \cite{jeongMoon_2013_selfiteratingSoftEqualizer}, as the latter are more sensitive to the convergence errors committed by the SISO decoder.
Block transmission in Proakis C channel is considered with $K=256$ and with QPSK, 8-PSK and 16-QAM constellations.
In Fig.~\ref{fig_ber_uncode}, BER of TV DFE with APP and EP feedback are compared to the proposed predictive IV implementations.
IV DFE converges towards the curve of TV counterparts, especially at high SNR, but it is seen that a gap remains at medium BER for some constellations, due to dynamic filtering capabilities of TV receivers.
EP feedback is shown to be mostly equivalent to APP feedback in this uncoded use case, but at high BER, EP has an advantage over APP both for TV and IV receivers, which suggests that improved decoding thresholds can be obtained with channel coding.
\subsection{On the Operating Regions of FIR Receivers}
A previous work on TV FIR turbo equalizers concluded that TV DFE significantly outperforms TV LE at high data rates \cite{sahinCipriano_2018_DFEEP}, whereas TV LE remains preferable at very low rates, as it achieves same performance with less complexity.
In the following, the asymptotic behaviour and the computational complexity of the proposed receiver along with IV FIRs is evaluated in a similar manner.
\begin{figure}[t]
\centering
\includegraphics[width=0.44\textwidth]{fig_8_asymptotic.pdf}
%\includegraphics[width=2.9in]{fig8_asymptotic.eps}
\caption{Achievable rates of FIR receivers for 8-PSK in Proakis C channel.
}
\label{fig_asymp}
\end{figure}
Through the extrinsic information transfer (EXIT) analysis of a SISO module, a mutual information (MI) based transfer function model, $I_E=\mathcal{T}_R(I_A, \mathbf{h}, \sigma_w^2)$ is obtained \cite{brink_2000_designing}, where $I_A$ and $I_E$ denote respectively the MI between coded bits and the prior LLRs and the extrinsic LLRs.
EXIT functions notably allow to numerically predict the achievable rates of SISO receivers, through the area theorem of EXIT charts \cite{hagenauer_exit_2004}.
Indeed, MAP detector's EXIT chart's area yields an accurate prediction of the channel symmetric information rate (SIR) \cite{arnold_simulation-based_2006}, the highest possible transmission rate for practical constellations, without channel knowledge at the transmitter.
However, for approximate receivers which violate the extrinsic message principle of turbo detection, the rates predicted by EXIT can be too optimistic (e.g. SIR of APP DFE appears to surpass MAP, which is impossible).
This has been observed for APP-based receivers in \cite{sahinCipriano_2018_DFEEP,liWuTao_2019_perfAnalysisAndImprovemetForVAMPFDE}, but the EP-based DFE does not suffer from this phenomenon.
Hence in the following, the proposed predictive EP-based IV DFE is evaluated.
IV DFE EP with symbol-wise prediction scheme is used for 8-PSK transmissions in the Proakis C channel, and numerically obtained achievable rates are plotted in solid lines in Fig. \ref{fig_asymp}.
Dotted plots illustrate the achievable rates without turbo-iterations, for each receiver.
IV receivers are shown to follow the behaviour of their TV counterpart within a gap of about $0.1$~bits/s/Hz for both LE and DFE, but IV DFE still keep a significant upper hand over TV LE at medium and high spectral efficiency operating points.
Using IV FIR receivers to operate at a given rate requires about $1.5$~dB more energy than TV FIR, but with significant complexity savings.
\begin{figure}[t]
\centering
\includegraphics[width=0.46\textwidth]{fig_9_coded_ber_pred_comp.pdf}
%\includegraphics[width=3in]{fig7_coded_ber_pred_comp.eps}
\caption{Rate-1/2 coded BER with proposed binary and symbol-wise prediction.
}
\label{fig_9_coded_ber_pred_comp}
\end{figure}
\begin{table}[!h]
\footnotesize
\renewcommand{\arraystretch}{1.2}
\caption{Computational Complexity of FIR Receivers}
\label{table_fir_complexity}
\centering
\begin{tabular}{c||c|c}
\hline
\bfseries Structure & \bfseries Filter Computation & \bfseries Filtering and Detection \\
\hline\hline
% backward recursion:
% N_\text{state}( N_\text{in} A_r + N_\text{in} (Log-MAP))
% = N_\text{state} (N_\text{in} A_r + N_\text{in} (I_r+2A_r))
%
\specialcell{TV LE} & \specialcell{$K(5L^3+56L^2)$} & $K(25L+(11+3q)M)$\\
\hline
\specialcell{TV DFE} & \specialcell{$K(5L^3+71L^2)$} & $K(25L+(18+3q)M)$\\
\hline
\specialcell{IV LE} & \specialcell{$(6L^3+28L^2)$} & $K(25L+(11+3q)M)$\\
\hline
\specialcell{IV DFE} & \specialcell{$(N_\text{pred}+1)(6L^3+34L^2)$} & $K(25L+(18+3q)M)$\\
\hline
\end{tabular}
\end{table}
\begin{figure*}[t]
\centering
\includegraphics[width=0.95\textwidth]{fig_10_coded_ber_all_rates.pdf}
%\includegraphics[width=7in]{fig9_coded_ber_all_rates.eps}
\caption{Coded 8-PSK bit-error rate (BER) performance comparison of turbo FIR receivers across turbo-iterations for different code rates.}
\label{fig_10_coded_ber_all_rates}
\end{figure*}
Approximate computational complexity per turbo-iteration of considered FIR receivers is given in the Table~\ref{table_fir_complexity}.
TV LE and DFE receivers use the reduced-complexity TV matrix inversion algorithms in \cite{sahinCipriano_2018_DFEEP}, and IV receivers exploit Cholesky decomposition for matrix inversion.
The filter computation cost of the proposed IV DFE increases linearly with the number of prediction iterations $N_\text{pred}$.
\subsection{Finite-Length Turbo-Equalization Performance}
In this section, the prediction accuracy is assessed for transmissions encoded with non-recursive non-systematic convolutional code (NRNSCC) of polynomial $[7,5]_8$.
First, the impact of choosing a symbol-wise or a binary prediction scheme is assessed through finite-length BER evaluations.
The block size is kept at $K=256$, similarly to the uncoded case, and a MAP decoder based on the BCJR algorithm is used as a SISO decoder \cite{bahl_optimal_1974}.
Fig.~\ref{fig_9_coded_ber_pred_comp} shows the case of the EP-based feedback with 8-PSK, and the use of symbol-wise prediction is shown to accelerate convergence of the IV DFE performance towards TV DFE.
However, despite the improvements brought by the symbol-wise prediction, covariance estimations tend to be too optimistic for high prior information at high SNRs (following 1 or 2 turbo iterations, above 15~dB), and degrade BER performance.
A similar observation was made for the semi-analytic prediction of turbo linear MMSE receivers in \cite{ning_extrinsic_2012}, where a calibration mechanism is applied to correct the predicted prior covariance with a multiplicative penalty factor.
After some ad hoc optimization, this scheme yields more pessimistic predictions that ends up improving the BER prediction accuracy.
Here, a related mechanism is adapted to the proposed online prediction.
To avoid over-estimation of the causal covariance, the anti-causal covariance can be exploited to derive a ``lower-bound'' to estimated causal covariances.
Empirically, turbo detection systems bring most of the improvements at the initial iterations, hence the improvements after a certain number of iterations can no longer be substantial.
Thus, after some trial-and-error tests, we have selected the predicted causal covariance $\bar{v}^c_x$ of the current turbo iteration to be modified with the heuristic $\bar{v}^c_x = \max(\bar{v}^c_x, \beta\bar{v}^a_x)$, with $0 < \beta \le 1$.
\begin{figure}[t]
\centering
\includegraphics[width=0.48\textwidth]{fig_11_ber_per_turbo_iters.pdf}
%\includegraphics[width=2.9in]{fig8_asymptotic.eps}
\caption{Achievable rates of FIR receivers for 8-PSK in Proakis C channel.
}
\label{fig_11_ber_per_turbo_iters}
\end{figure}
The proposed heuristic is integrated with the symbol-wise prediction, with 3 prediction iterations and $\beta=0.2$, and the IV DFE-EP performance is displayed in turquoise in Fig. \ref{fig_9_coded_ber_pred_comp}.
Finally, to compare our proposal to the prior work and to evaluate its behavior in different operating regimes, the previously used rate-1/2 encoding with NRNSCC $[7,5]_8$ is punctured to get rate-2/3 encoding with $[1 1; 0 1]$ puncturing pattern and rate-5/6 encoding with $[1 0 0 0 1; 0 1 1 1 1]$ puncturing pattern.
The BER performance of the proposed IV DFE APP and IV DFE EP receivers are shown in red in Fig.~\ref{fig_10_coded_ber_all_rates}, for 8-PSK transmissions in Proakis C channel, with above mentioned codes of rate 1/2, 2/3 and 5/6, and for 0, 1 and 4 turbo-iterations.
Proposed predictive IV DFE receivers use symbol-wise prediction with 3 iterations, and the heuristic parameter is $\beta=0.2$.
IV DFE APP significantly outperforms other APP-based DFE receivers when there are no turbo iterations, as this is the operating point where the prediction scheme is the most accurate.
In Figure~\ref{fig_11_ber_per_turbo_iters}, the evolution of BER is plotted as the number of turbo-iterations increases.
During intermediary iterations of the rate $2/3$ system, previous works of \emph{Tao et al.} \cite{tao_2016_low} and \emph{Lou et al.} \cite{lou_soft_2014} close most of the gap of iteration zero, with the receiver of \cite{lou_soft_2014} slowly converging to the same limit as the proposed receiver.
At high rate systems (rate 5/6) the gap between them and our proposal increases, even for 4 turbo-iterations, and from Figure~\ref{fig_11_ber_per_turbo_iters} it is seen that the receiver of \cite{tao_2016_low} cannot converge to the same asymptotic limits, probably due to the usage of only a few samples for covariance estimation heuristic.
The use of {EP}-feedback instead of {APP} does not bring significant improvement for high-rates, or without turbo-iterations, but at medium and low rates, it allows for an additional asymptotic gain over 0.5~dB.
However, the predictability of the {EP} feedback over a wider set of configurations makes it a more attractive solution.
\section{Conclusion}
This paper carries out an original approach to the design of turbo DFE receivers with static filters, through the use of online prediction, based on semi-analytic performance prediction techniques as used in physical layer abstraction methods.
Due to the lack of a closed-form solution for such receivers, various heuristics are used throughout the literature. However, discussion on the optimality of such approaches was lacking and it is one of the contribution of this paper.
Here, semi-analytical performance prediction of exact time-varying turbo DFE with dynamic filters is exploited to derive static DFE filters.
This approach has been carried out for DFE with APP-based or EP-based soft feedback and their detection performance has been evaluated in various configurations.
This framework could also be applied to self-iterated FIR DFE \cite{sahinCipriano_2018_DFEEP} for further improved performance, by updating anti-causal variances with causal EP variance of the last self-iteration.
Our analysis shows that significant complexity savings can be achieved with respect to TV DFE, while offering reasonably close performance.
Moreover, our method is compatible with any constellation, and spectrally efficient on a large interval of coding rates and with or without turbo-iterations.
% Can use something like this to put references on a page
% by themselves when using endfloat and the captionsoff option.
\ifCLASSOPTIONcaptionsoff
\newpage
\fi
% trigger a \newpage just before the given reference
% number - used to balance the columns on the last page
% adjust value as needed - may need to be readjusted if
% the document is modified later
%\IEEEtriggeratref{8}
% The "triggered" command can be changed if desired:
%\IEEEtriggercmd{\enlargethispage{-5in}}
% references section
% can use a bibliography generated by BibTeX as a .bbl file
% BibTeX documentation can be easily obtained at:
% http://mirror.ctan.org/biblio/bibtex/contrib/doc/
% The IEEEtran BibTeX style support page is at:
% http://www.michaelshell.org/tex/ieeetran/bibtex/
%\bibliographystyle{IEEEtran}
% argument is your BibTeX string definitions and bibliography database(s)
%\bibliography{IEEEabrv,../bib/paper}
\bibliographystyle{IEEEtran}
\bibliography{IEEEabrv,bibDFE}
%
% <OR> manually copy in the resultant .bbl file
% set second argument of \begin to the number of references
% (used to reserve space for the reference number labels box)
%\begin{thebibliography}{1}
%\bibitem{IEEEhowto:kopka}
%H.~Kopka and P.~W. Daly, \emph{A Guide to \LaTeX}, 3rd~ed.\hskip 1em plus
% 0.5em minus 0.4em\relax Harlow, England: Addison-Wesley, 1999.
%\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}
\ifdouble
\begin{IEEEbiography}[{\includegraphics[width=1in,height=1.25in,clip,keepaspectratio]{sahin.jpg}}]{Serdar \c{S}ahin}
was born in Ankara, Turkey, in 1992.
He received the M.Sc.Eng. degree in control sys-tems and electronics engineering from INSA de Toulouse, University of Toulouse, France, in 2015.
He is currently pursuing the PhD degree in digital communications with IRIT-ENSEEIHT, Toulouse, and also with Thales Communications and Security, Gennevilliers.
His main research interests include iterative receiver design, practical cooperative transmission schemes and PHY layer abstraction.
\end{IEEEbiography}
\begin{IEEEbiography}[{\includegraphics[width=1in,height=1.25in,clip,keepaspectratio]{cipri.jpg}}]{Antonio Maria Cipriano} was born in Padova, Italy, in 1976. He received the Laurea degree in telecommunications engineering from the University of Padova, Italy, in 2000, and the Ph.D. degree in digital communications jointly from the University of Padova and the Ecole Nationale Sup\'{e}rieure des T\'{e}l\'{e}communications (ENST) Paris, France, in 2005.
In 2001, he was a Young Engineer at Eutelsat, France, for eight months.
From 2005 to 2007, he held two post-doctoral positions at ENST, Paris, and Orange Labs.
In 2007 he joined Thales Communication and Security as a digital communication engineer and was involved in several national and international research projects on 4G and 5G communication systems.
His main research interests lie in the broad area of digital communication systems.
He is currently involved in research about PHY layer abstractions, relaying for ad hoc mobile networks and advanced receiver design.
\end{IEEEbiography}
\begin{IEEEbiography}[{\includegraphics[width=1in,height=1.25in,clip,keepaspectratio]{poull.jpg}}]{Charly Poulliat}
received the M.Sc.Eng. degree in electrical engineering from the Ecole Nationale Sup\'{e}rieure de l'Electronique et de ses Applications (ENSEA), Cergy-Pontoise, France, in 2001 and the M.S. degree in Image and Signal Processing from the University of Cergy-Pontoise, France, in 2001, and his PhD degree in electrical and computer engineering from the University of
Cergy-Pontoise, France, in 2004, and the Habilitation degree from the University of Cergy-Pontoise in 2010.
From 2004 to 2005, he was a Post-Doctoral Researcher at UH Coding Group, University of Hawaii at Manoa, HI, USA, supervised by Prof. M. Fossorier.
In 2005, he joined the Signal and Telecommunications department of the engineering school ENSEA as an Assistant Professor.
Since 2011, he has been a Professor with the
National Polytechnic Institute of Toulouse (INP-ENSEEIHT), University of Toulouse. He is also with the Signal and Communications Group, CNRS IRIT Laboratory.
His research interests include signal processing for digital communications, waveform design, channel coding,
iterative system design and optimization.
\end{IEEEbiography}
\begin{IEEEbiography}[{\includegraphics[width=1in,height=1.25in,clip,keepaspectratio]{bouch.jpg}}]{Marie-Laure Boucheret}
received the M.Sc.Eng. degree in electrical engineering from ENST Bretagne, Brest, France in 1985, the Ph.D. degree in Digital Communications from TELECOM ParisTech, in 1997, and the Habilitation \`{a} diriger les recherches degree from INPT University of Toulouse, in 1999.
From 1985 to 1986, she was a Research Engineer with the French Philips Research Laboratory (LEP). From 1986 to 1991, she was an Engineer with Thales Alenia Space, first as a Project Engineer (TELECOM II program) then as a Study Engineer at the Transmission Laboratory.
From 1991 to 2005, she was with TELECOM ParisTech first as an Associated Professor then as a Professor.
Since 2005, she has been a Professor with the National Polytechnic Institute of Toulouse (INP-ENSEEIHT),
University of Toulouse. She is also with the Signal and Communication Group, IRIT Laboratory. Her fields of interest are signal processing for communication and satellite communications.
\end{IEEEbiography}
\else
\fi
% that's all folks
\end{document}
% Generated by IEEEtran.bst, version: 1.14 (2015/08/26)
\begin{thebibliography}{10}
\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{sahinCipriano_2018_DFEEP}
S.~\c{S}ahin, A.~M. Cipriano, C.~Poulliat, and M.~Boucheret, ``Iterative
equalization with decision feedback based on expectation propagation,''
\emph{{IEEE} Trans. Commun.}, vol.~66, no.~10, pp. 4473--4487, Oct. 2018.
\bibitem{tuchler_turbo_2011}
M.~T\"{u}chler and A.~C. Singer, ``Turbo {Equalization}: {An} {Overview},''
\emph{{IEEE} Trans. Inf. Theory}, vol.~57, no.~2, pp. 920--952, Feb. 2011.
\bibitem{lou_soft_2014}
H.~Lou and C.~Xiao, ``The soft-feedback {ISI} canceller-based turbo equalizer
for multilevel modulations,'' \emph{Int. J. of Wireless Inform. Networks},
vol.~21, no.~1, pp. 68--73, Mar. 2014.
\bibitem{tao_2016_low}
J.~Tao, ``On low-complexity soft-input soft-output decision-feedback
equalizers,'' \emph{{IEEE} Comm. Let.}, vol.~20, no.~9, pp. 1737--1740, Sep.
2016.
\bibitem{jeong_2011_low_complexity_bidir}
S.~Jeong, ``Low complexity turbo equalizations and lower bounds on information
rate for intersymbol interference channels,'' Ph.D. dissertation, University
of Minnesota, Oct. 2011.
\bibitem{belfiore_79_DFE}
C.~A. Belfiore \emph{et~al.}, ``Decision feedback equalization,'' \emph{Proc.
of the IEEE}, vol.~67, no.~8, pp. 1143--1156, Aug. 1979.
\bibitem{tuchler_turbo_2002}
M.~T\"{u}chler \emph{et~al.}, ``Turbo equalization: principles and new
results,'' \emph{{IEEE} Trans. Commun.}, vol.~50, no.~5, pp. 754--767, May
2002.
\bibitem{jeong_turbo_2010}
S.~Jeong and J.~Moon, ``Turbo {Equalization} {Based} on {Bi}-{Directional}
{DFE},'' in \emph{{IEEE} {Int.} {Conf.} on {Commun.}}, May 2010, pp. 1--6.
\bibitem{lopes_soft-feedback_2006}
R.~R. Lopes and J.~R. Barry, ``The soft-feedback equalizer for turbo
equalization of highly dispersive channels,'' \emph{{IEEE} Trans. Commun.},
vol.~54, no.~5, pp. 783--788, May 2006.
\bibitem{minka_2001_expectationpropagation}
T.~P. Minka, ``A family of algorithms for approximate bayesian inference,''
{Ph.D.} dissertation, M.I.T., Jan. 2001.
\bibitem{liWuTao_2019_perfAnalysisAndImprovemetForVAMPFDE}
D.~Li, Y.~Wu, J.~Tao, and M.~Zhu, ``Performance analysis and improvement for
vamp soft frequency-domain equalizers,'' \emph{IEEE Access}, vol.~7, pp.
42\,495--42\,506, Mar. 2019.
\bibitem{balakrishnan_mitigation_1999}
J.~Balakrishnan, ``Mitigation of error propagation in decision feedback
equalization,'' Master's {Thesis}, Cornell Univ., Urbana-Champaign, 1999.
\bibitem{visoz_semi-analytical_2010}
R.~Visoz, A.~O. Berthet \emph{et~al.}, ``Semi-analytical performance prediction
methods for iterative {MMSE}-{IC} multiuser {MIMO} joint decoding,''
\emph{{IEEE} Trans. Commun.}, vol.~58, no.~9, pp. 2576--2589, Sep. 2010.
\bibitem{sabbaghian_2009_analyticalapproach}
M.~Sabbaghian and D.~Falconer, ``An analytical approach for finite block length
performance analysis of turbo frequency-domain equalization,'' \emph{{IEEE}
Trans. Veh. Technol.}, vol.~58, no.~3, pp. 1292--1301, Mar. 2009.
\bibitem{yuan_2008_evolutionanalysis}
X.~Yuan, Q.~Guo, X.~Wang, and L.~Ping, ``Evolution analysis of low-cost
iterative equalization in coded linear systems with cyclic prefixes,''
\emph{{IEEE} J. Sel. Areas Commun.}, vol.~26, no.~2, pp. 301--310, Feb. 2008.
\bibitem{brink_2000_designing}
S.~Ten~Brink, ``Designing iterative decoding schemes with the extrinsic
information transfer chart,'' in \emph{AEU Int. J. Electron. Commun.},
vol.~54, no.~6, Jan. 2000.
\bibitem{fu_2005_stochasticanal}
M.~Fu, ``Stochastic analysis of turbo decoding,'' \emph{{IEEE} Trans. Inf.
Theory}, vol.~51, no.~1, pp. 81--100, Jan. 2005.
\bibitem{ibing_2010_predictingconvergence}
A.~Ibing and H.~Boche, ``On predicting convergence of iterative {MIMO}
detection-decoding with concatenated codes,'' \emph{{IEEE} Trans. Veh.
Technol.}, vol.~59, no.~8, pp. 4134--4139, Oct. 2010.
\bibitem{jeongMoon_2013_selfiteratingSoftEqualizer}
S.~Jeong and J.~Moon, ``Self-iterating soft equalizer,'' \emph{{IEEE} Trans.
Commun.}, vol.~61, no.~9, pp. 3697--3709, Sep. 2013.
\bibitem{hagenauer_exit_2004}
J.~Hagenauer, ``The {EXIT} chart - introduction to extrinsic information
transfer in iterative processing,'' in \emph{Proc. {IEEE} 12th European
Signal Process. Conf.}, Sep. 2004, pp. 1541--1548.
\bibitem{arnold_simulation-based_2006}
D.~M. Arnold, H.~A. Loeliger \emph{et~al.}, ``Simulation-based computation of
information rates for channels with memory,'' \emph{{IEEE} Trans. Inf.
Theory}, vol.~52, no.~8, pp. 3498--3508, Aug. 2006.
\bibitem{bahl_optimal_1974}
L.~Bahl, J.~Cocke, F.~Jelinek, and J.~Raviv, ``Optimal decoding of linear codes
for minimizing symbol error rate,'' \emph{{IEEE} Trans. Inf. Theory},
vol.~20, no.~2, pp. 284--287, Mar. 1974.
\bibitem{ning_extrinsic_2012}
B.~Ning, R.~Visoz, and A.~O. Berthet, ``Extrinsic versus a posteriori
probability based iterative {LMMSE}-{IC} algorithms for coded {MIMO}
communications: {Performance} and analysis,'' in \emph{Proc. {IEEE} Int.
Symp. on Wireless Commun. Systems}, Aug. 2012, pp. 386--390.
\end{thebibliography}