Tight Upper Bounds on the Error Probability of Spinal Codes over Fading Channels
Extracted source text · arXiv 2304.13910
%\documentclass[journal,draftclsnofoot,onecolumn]{IEEEtran}
%\documentclass[journal]{IEEEtran}
\documentclass[conference,a4paper]{IEEEtran}
\IEEEoverridecommandlockouts
\addtolength{\topmargin}{9mm}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage{url} % provides \url{...}
%\usepackage{ifthen} % provides \ifthenelse
\usepackage{cite} % improves presentation of citations
\usepackage[cmex10]{amsmath} % Use the [cmex10] option to ensure complicance
% with IEEEXplore (see bare_conf.tex)
\interdisplaylinepenalty=1000 % As explained in bare_conf.tex
\usepackage{mleftright} % fix to wrong spacing of \left-,
\mleftright % \middle- \right-commands
\usepackage{booktabs} % fixes poor spacing in tables and
% provides \toprule, \midrule, \bottomrule
\hyphenation{op-tical net-works semi-conduc-tor}
%\usepackage[explicit]{titlesec}%更改标题间距
%\usepackage[title]{appendix}
\usepackage{verbatim}
\usepackage{graphicx}
\usepackage{color, soul}
\usepackage{epstopdf}
\usepackage{subfigure} % Written by Steven Douglas Cochran
%\usepackage{algorithm}
\usepackage{algorithmic}
%\usepackage{amsmath}
\usepackage{amsthm}
\usepackage{extarrows}
\usepackage{mathrsfs}
\newtheorem{theorem}{Theorem}
\newtheorem{example}{Example}
\newtheorem{case}{Case}
\usepackage{bbm}
%\usepackage{geometry}
\usepackage[backref]{hyperref}
\usepackage{amssymb}
\newtheorem{remark}{Remark}
\newtheorem{property}{Property}
\newcommand{\RNum}[1]{\uppercase\expandafter{\romannumeral #1\relax}}
\newtheorem{lemma}{Lemma}
\newtheorem{definition}{Definition}
\newtheorem{corollary}{Corollary}
\usepackage{stfloats}
\newtheorem{problem}{Problem}
\usepackage{color}
\usepackage{tikz,xcolor,hyperref}% Make Orcid icon, 并在参考文献中加上返回值
\newcommand{\E}{\mathbb{E}}
%\usepackage{flushend}
\definecolor{lime}{HTML}{A6CE39}
% Define the ORCID iD command for each author separately. Here done for two authors.
%两个引用直接没有括号
%\usepackage[numbers,sort&compress]{natbib}
%\titlespacing{\section}{0pt}{1.2ex plus .0ex minus .0ex}{.3ex plus .0ex}%更改标题间距
%\titlespacing{\subsection}{0pt}{1.2ex plus .0ex minus .0ex}{.3ex plus .0ex}%更改标题间距
\allowdisplaybreaks[4]
\makeatletter
\DeclareRobustCommand{\orcidicon}{%
\begin{tikzpicture}
\draw[lime, fill=lime] (0,0)
circle [radius=0.16]
node[white] {{\fontfamily{qag}\selectfont \tiny ID}}; \draw[white, fill=white] (-0.0625,0.095)
circle [radius=0.007]; \end{tikzpicture}
\hspace{-2mm}}
\foreach \x in {A, ..., Z}{%
\expandafter\xdef\csname orcid\x\endcsname{\noexpand\href{https://orcid.org/\csname orcidauthor\x\endcsname}{\noexpand\orcidicon}}
}
\newcommand*\bigcdot{\mathpalette\bigcdot@{.5}}
\newcommand*\bigcdot@[2]{\mathbin{\vcenter{\hbox{\scalebox{#2}{$\m@th#1\bullet$}}}}}
\makeatother
%\usepackage{geometry}
%\usepackage{stfloats}
\let\algorithm\relax
\let\endalgorithm\relax
\usepackage[linesnumbered,ruled,vlined]{algorithm2e}%[ruled,vlined]{
\renewcommand{\algorithmicrequire}{\textbf{Input:}} % Use Input in the format of Algorithm
\renewcommand{\algorithmicensure}{\textbf{Output:}} % Use Output in the format of Algorithm
\renewcommand{\algorithmicrequire}{\textbf{Input:}} % Use Input in the format of Algorithm
\renewcommand{\algorithmicensure}{\textbf{Output:}} % Use Output in the format of Algorithm
\newcommand{\supercite}[1]{\textsuperscript{\cite{#1}}}
%\usepackage{float}%图片位置
\hyphenation{op-tical net-works semi-conduc-tor}
\raggedbottom
\makeatletter
\renewcommand{\maketag@@@}[1]{\hbox{\m@th\normalsize\normalfont#1}}%
\makeatother
\begin{document}
\title{Tight Upper Bounds on the Error Probability of Spinal Codes over Fading Channels}
\author{%
\IEEEauthorblockN{Xiaomeng~Chen, Aimin~Li and Shaohua~Wu}
\IEEEauthorblockA{Harbin Institute of Technology (Shenzhen)\\
518055 Shenzhen, China\\
cxmeng@163.com, liaimin@stu.hit.edu.cn, hitwush@hit.edu.cn}
% \and
% \IEEEauthorblockN{Aimin~Li}
% \IEEEauthorblockA{Harbin Institute of Technology (Shenzhen)\\
% 518055 Shenzhen, China\\
% liaimin@stu.hit.edu.cn}
% \and
% \IEEEauthorblockN{Shaohua~Wu}
% \IEEEauthorblockA{Harbin Institute of Technology (Shenzhen)\\
% 518055 Shenzhen, China\\
% hitwush@hit.edu.cn}
\thanks{
Aimin Li has contributed equally to this work. This work has been supported in part by the National Key Research and Development Program of China under Grant no. 2020YFB1806403, and in part by the National Natural Science Foundation of China under Grant nos. 61871147, 62071141, 61831008, 62027802, and in part by the Shenzhen Municipal Science and Technology Plan under Grant no. GXWD20201230155427003-20200730122528002, and in part by the Major Key Project of PCL under Grant PCL2021A03-1. (Corresponding author: Shaohua Wu.)
}
}
% make the title area
\maketitle
% \allowdisplaybreaks
% As a general rule, do not put math, special symbols or citations
% in the abstract or keywords.
\begin{abstract}
Spinal codes, a family of rateless codes introduced in 2011, have been proved to achieve Shannon capacity over both the additive white Gaussian noise (AWGN) channel and the binary symmetric channel (BSC). In this paper, we derive explicit tight upper bounds on the error probability of Spinal codes under maximum-likelihood (ML) decoding and perfect channel state information (CSI) over three typical fading channels, including the Rayleigh channel, the Nakagami-m channel and the Rician channel. Simulation results verify the tightness of the derived upper bounds.
\end{abstract}
% Note that keywords are not normally used for peerreview papers.
\begin{IEEEkeywords}
%Semantic communications, age of information, goal-oriented communications, effectiveness of information.
Spinal codes, decoding error probability, fading channels, ML decoding. upper bounds.
\end{IEEEkeywords}
\IEEEpeerreviewmaketitle
\section{Introduction}
First proposed in 2011 \cite{2011Spinal}, Spinal codes are a family of rateless codes that have been proved to achieve Shannon capacity over both the AWGN channel and the BSC \cite{balakrishnan2012randomizing}. Possessing the rateless capacity-achieving nature, Spinal codes demonstrate their superiority in bridging a reliable and high-efficiency information delivery pipeline between transceivers under highly dynamic channels.
Specifically, the rateless nature allows Spinal codes to automatically adapt to time-varying channel conditions. Unlike fixed-rate codes, which require a specific code rate in advance, rateless codes work by a natural channel adaptation manner: the sender transmits a potentially limitless stream of encoded bits, and the receiver collects bits consecutively until the successful decoding process takes place.
In \cite{2012Spinal}, Spinal codes have shown advantages in terms of rate performance compared to other state-of-the-art rateless codes %, including Raptor codes \cite{Raptor,RCONC}, Strider codes \cite{Strider} and rateless LDPC codes \cite{LDPC},
under different channel conditions and message sizes.
Also, \cite{2012Spinal} notes the similarity between Spinal codes and Trellis Coded Modulation (TCM) \cite{TWC1,TWC2} is superficial because of their differences in nature and encoded popuse.
In coding theory, performance analysis is an intriguing topic. A closed-form expression for the error probability, in general, could not only enable more efficient performance evaluations but also shed light on coding optimization design.
However, in most cases, it is intractable to obtain a closed-form expression for error probabilities.
As an alternative, bounding techniques are usually used to approximate performance \cite{NBLLBC}.
Along this avenue, there are already a lot of established bounds for some specific channel codes, such as the advanced tight bounds on Polar codes \cite{polar1,polar2}, the upper and lower bounds on Raptor codes \cite{raptorerror}, and the performance bounds on the LT codes \cite{LTerror1,LTerror2}.
And in \cite{CIT-009}, upper and lower bounds on the error probability of linear codes under ML decoding are surveyed.
For Spinal codes, however, it is still early days to get tight explicit bounds over fading channels.
To date, there have been a few works that have preliminarily explored the performance analysis of Spinal codes. In \cite{balakrishnan2012randomizing}, Balakrishnan \emph{et al.} analyze the asymptotic rate performance of Spinal codes and theoretically prove that Spinal codes are capacity-achieving over both the AWGN channel and the BSC. In \cite{UEPspinal}, two state-of-the-art results in the finite block length (FBL) regime, \emph{i.e.}, the Random Coding Union (RCU) bound \cite{polyanskiy2010channel} and a variant of Gallager result \cite{gallager1968information} are applied to analyze the error probability of Spinal codes over the AWGN channel and the BSC, respectively. In \cite{li2021new}, the authors further tighten bounds by characterizing the error probability as the volume of a hypersphere divided by the volume of a hypercube, while the analysis is still performed over the AWGN channel. Until now, little has been done in the way of error probability analysis for Spinel codes over fading channels. An exception work is \cite{li2021spinal}, which derived a probability-dependent convergent upper bound over Rayleigh fading channels by adopting the Chernoff bound \cite{chernoff1952measure}. However, $i$) the derived bound over the Rayleigh fading channel in \cite{li2021spinal} is not strictly explicit, since the convergence of the upper bound is probability-dependent; $ii$) Nakagami-m and Rician channels, both common in practical wireless communication scenarios, have not been considered in the available error probability analyses; $iii$) There is not yet an upper bound that achieves uniform tight approximations over a wide range of signal-to-noise ratio (SNR), either over the fading channel or over the AWGN channel.
Motivated by the above, this paper aims to derive tight upper bounds over three typical fading channels, Rayleigh, Nakagami-m and Rician fading channels. Strictly explicit upper bounds are provided, which cover uniform tight error probability approximations over a wide range of SNR.
The paper is organized as follows. In Section \ref{section II}, the encoding process of Spinal codes is given as a priori knowledge. Section \ref{section III} derives upper bounds over Rayleigh, Nakagami-m and
\begin{figure}
\centering
\includegraphics[width= 0.9\linewidth]{figure1plus.pdf}
\caption{The encoding process of Spinal codes.}
\label{figure1}
\end{figure}
Rician fading channels, respectively. Numerical results comparing the bounds with Monte Carlo simulations are illustrated in Section \ref{sectionIV}. Section \ref{sectionV} presents the conclusions of our work.
\section{Encoding Process of Spinal Codes} \label{section II}
This section briefly introduces the primary building blocks of Spinal codes, where the combination of a hash function and random number generator (RNG) functions is a key. %The properties that the hash function fulfills can be seen in Appendix A, which paves the way for the performance analysis.
Fig. \ref{figure1} shows the encoding process of Spinal codes, which comprises five steps:
\begin{enumerate}
\item \textbf{Segmentation}: Divide an $n$-bit message $\mathbf{M}$ into $k$-bit segments $\mathbf{m}_i \in \left\{0,1 \right\}^k$, where $i=1,2,\cdots,n/k$.
\item \textbf{Sequentially Hashing}\footnote{A discussion concerning the properties of the hash function can be found in Appendix \ref{property}.}: The hash function sequentially generates $v$-bit spine values $\mathbf{s}_i \in {\{0,1\}}^v$, with
\begin{equation} \label{eqhash}
\mathbf{s}_i = h(\mathbf{s}_{i-1},\mathbf{m}_i),~i=1,2,\cdots,n/k,~\mathbf{s}_0 = \mathbf{0}^v \text{.}
\footnote{The initial spine value $\mathbf{s}_0$ is known to both the encoder and the decoder, and is set as $\mathbf{s}_0=0$ in this papser without loss of generality.}
\end{equation}
\item \textbf{RNG}: Each spine value $\mathbf{s}_i$ is used to seed an RNG to generate a binary pseudo-random uniform-distributed sequence $\{\mathbf{x}_{i,j}\}_{j \in \mathbb{N}^+}$:
\begin{equation}
\mathrm{RNG:} ~\mathbf{s}_i \to \{ \mathbf{x}_{i,j} \},~j=1,2,3,\cdots \text{,}
\end{equation}
where $\mathbf{x}_{i,j} \in {\{ 0,1 \}}^c$.
\item \textbf{Constellation Mapping}: The constellation mapper maps each $c$-bit symbol $\mathbf{x}_{i,j}$ to a channel input set $\Psi$:
\begin{equation}
\operatorname{f}: \mathbf{x}_{i,j}\rightarrow \Psi \text{.}
\end{equation}
In this paper, $\operatorname{f}$ is a uniform constellation mapping function, \emph{i.e.}, it converts each $c$-bit symbol $\mathbf{x}_{i,j}$ to a decimal symbol $\operatorname{f}(\mathbf{x}_{i,j})$.
\item \textbf{Rateless Transmission}: The encoder continues to generate and transmit symbols pass by pass until the receiver successfully decodes the message.
\end{enumerate}
\section{UPPER BOUNDS ON THE ERROR PROBABILITY}\label{section III}
Error probability analysis is of primary importance in coding theory.
Closed-form expressions for error probabilities, however, are intractable in most cases. To address this issue, a commonly adopted alternative approach is to introduce bounding techniques for the error probability approximation. In line with this idea, this section aims to derive tight upper bounds on the error probability for Spinal codes over Rayleigh, Nakagami-m and Rician fading channels, respectively.
\subsection{Upper Bounds on the Rayleigh Fading Channel}
\begin{theorem} \label{theorem1}
Consider Spinal codes with message length $n$, segmentation parameter $k$, modulation parameter $c$, and sufficiently large hash parameter $v$ transmitted over a flat slow Rayleigh fading channel with mean square $\Omega$ and AWGN variance $\sigma^2$, the average error probability given perfect channel state information (CSI) under ML decoding for Spinal codes can be upper bounded by
\begin{equation} \label{eq1}
\small
P_e \leq 1-\prod_{a=1}^{n/k} \left(1-\epsilon_a\right) \text{,}
\end{equation}
with
\begin{equation} \label{eq2}
\epsilon_a = \mathrm{min} \left\{ 1,\left(2^k-1\right)2^{n-ak} \cdot \mathscr{F}_{\text{Rayleigh}}(L_a,\sigma) \right\} \text{,}
\end{equation}
\begin{equation} \label{eq3}
\small
\mathscr{F}_{\text{Rayleigh}} \left(L_a \text{,} \sigma \right) = \sum_{r=1}^{N} b_r\mathcal{F}_{\text{Rayleigh}}(\theta_r \text{;} \sigma\text{,}L_a)\text{,}
\end{equation}
\begin{equation} \label{eq4}
\small
\mathcal{F}_{\text{Rayleigh}}(\theta_r\text{;} \sigma\text{,}L_a)=
{\left( \sum_{i \in \Psi} \sum_{j \in \Psi} 2^{-2c} \frac{8{\sigma}^2{\sin}^2{\theta}_r}{{\Omega}(i-j)^2+8{\sigma}^2{\sin}^2{\theta}_r} \right)}^{L_a}\text{,}
\end{equation}
where $b_r = \frac{\theta_r-\theta_{r-1}}{\pi}$, $L_a = (n/k-a+1)L$, $L$ is the number of transmitted passes, $\theta_r$ is arbitrarily chosen with $\theta_0=0$, $\theta_N=\frac{\pi}{2}$ and $0 < \theta_1 < \theta_2 < \cdots < \theta_{N-1} < \frac{\pi}{2}$, and $N$ represents the number of $\theta$ values which enables the adjustment of accuracy.
\end{theorem}
\begin{proof}
Suppose a message $ \mathbf{M}=\left(\mathbf{m}_1,\mathbf{m}_2,\cdots,\mathbf{m}_{n/k}\right)\in {\{0,1\}}^n $ is encoded to $\operatorname{f}\left(\mathbf{x}_{i,j}(\mathbf{M})\right)$ to be transmitted over a flat slow Rayleigh fading channel with an AWGN. The received signal can be expressed as:
\begin{equation} \label{eq5}
y_{i,j} = h_{i,j}\operatorname{f}\left(\mathbf{x}_{i,j}(\mathbf{M})\right) + n_{i,j} \text{,}
\end{equation}
where $y_{i,j}$ is the corresponding received signal at the receiver, $h_{i,j}$ is the channel fading parameter following Rayleigh distribution with mean square $\Omega$, and $n_{i,j}$ is the AWGN with variance $\sigma^2$.
ML decoding selects the one with the lowest decoding cost from the candidate sequence space ${\{0,1\}}^n$. Given received symbols $y_{i,j}$ and perfect CSI $\hat{h}_{i,j}=h_{i,j}$, the ML rule for the Rayleigh fading channel with the AWGN is
\begin{equation} \label{eq7}
\small
\begin{split}
\hat{\mathbf{M}} &\in \underset{\mathbf{\bar{M}} \in {\{0,1\}}^n}{\mathrm{arg\,min}} \sum_{i=1}^{n/k} \sum_{j=1}^{L} {(y_{i,j}-h_{i,j}\operatorname{f}(\mathbf{x}_{i,j}(\bar{\mathbf{M}})))}^2 \text{,}
\end{split}
\end{equation}
where $\hat{\mathbf{M}}$ is the decoding result, $\bar{\mathbf{M}}$ is the candidate sequence.
From this perspective, we classify the set of candidate sequences ${\{0,1\}}^n$ into two subsets: $i$) the correct decoding sequence $\mathbf{M} = \left(\mathbf{m}_1,\mathbf{m}_2,\cdots,\mathbf{m}_{n/k}\right)$; $ii$) wrong decoding sequences $\mathbf{M}' = \left(\mathbf{m}'_1,\mathbf{m}'_2,\cdots,\mathbf{m}'_{n/k}\right) \in \mathcal{W}$, with $\mathcal{W} \triangleq \left\{ \left(\mathbf{m}'_1,\mathbf{m}'_2,\cdots,\mathbf{m}'_{n/k}\right) : \exists 1 \leq i \leq n/k, \mathbf{m}'_i \neq \mathbf{m}_i \right\}$. Denoting the cost of $\mathbf{M}$ as $\mathscr{D}(\mathbf{M})$, it turns out that
\begin{align} \label{eq8}
\small
\begin{split}
\mathscr{D}(\mathbf{M}) &\triangleq \sum_{i=1}^{n/k}\sum_{j=1}^{L}{\left(y_{i,j}-h_{i,j}\operatorname{f}(\mathbf{x}_{i,j}(\mathbf{M}))\right)}^2 = \sum_{i=1}^{n/k}\sum_{j=1}^{L} n_{i,j}^2 \text{.}
\end{split}
\end{align}
Similarly, denote the cost of $\mathbf{M}'$ as $\mathscr{D}(\mathbf{M}')$, given by:
{
\small
\begin{align} \label{eq9}
\mathscr{D}(\mathbf{M}') \triangleq \sum_{i=1}^{n/k}\sum_{j=1}^{L}{\left(y_{i,j}-h_{i,j}\operatorname{f}(\mathbf{x}_{i,j}(\mathbf{M}'))\right)}^2 \text{.}
\end{align}
}
In the sequal, we attempt to explicitly express the error probability of Spinal codes. Let $\mathcal{E}_a$ be the event that there exists an error in the $a^{th}$ segment, which implies that:
\begin{enumerate}
\item The $a^{th}$ segment is different, \emph{i.e.}, $\mathbf{m}_a \neq \mathbf{m}'_a$.
\item The cost of the wrong decoding sequence is less than the correct one, \emph{i.e.}, $\mathscr{D}(\mathbf{M}') \leq \mathscr{D}(\mathbf{M})$. In this case, the ML decoder will incorrectly choose a certain wrong sequence $\mathbf{M}'\in \mathcal{W}$ as the decoding output.
\end{enumerate}
Denote $\overline{\mathcal{E}}_a$ as the complement of $\mathcal{E}_a$. The error probability of Spinal codes can be expressed as:
\begin{equation} \label{eq10}
\small
\begin{split}
P_e &= \mathrm{Pr}\left( \bigcup_{a=1}^{n/k}\mathcal{E}_a \right) = 1 - \mathrm{Pr}\left(\bigcap_{a=1}^{n/k} \overline{\mathcal{E}}_a \right) \\
%&= 1 - \prod_{a=1}^{n/k}\mathrm{Pr} \left( \overline{\mathcal{E}}_a\bigg|\bigcap_{i=1}^{a-1}\overline{\mathcal{E}}_i \right) \\
&= 1 - \prod_{a=1}^{n/k} \left[ 1 - \mathrm{Pr} \left( \mathcal{E}_a\bigg|\bigcap_{i=1}^{a-1}\overline{\mathcal{E}}_i \right)\right] \text{.}
\end{split}
\end{equation}
Thus, to obtain the error probability $P_e$, the remaining issue is to calculate $\mathrm{Pr} \bigg( \mathcal{E}_a\bigg|\bigcap_{i=1}^{a-1}\overline{\mathcal{E}}_i \bigg)$, which is interpreted as the probability that the $a^{th}$ segment is wrong while the previous $(a-1)$ segments are correct. With the defination that $\mathcal{W}_a \triangleq \{\left(\mathbf{m}'_1, \cdots ,\mathbf{m}'_a\right)\text{:} \mathbf{m}'_1=\mathbf{m}_1,\cdots,\mathbf{m}'_{a-1}=\mathbf{m}_{a-1},\mathbf{m}'_a \neq \mathbf{m}_a\} \subseteq \mathcal{W}$, the conditional probability is transformed as:
\begin{equation} \label{eq11}
\small
\mathrm{Pr} \left( \mathcal{E}_a\bigg|\bigcap_{i=1}^{a-1}\overline{\mathcal{E}}_i \right) = \mathrm{Pr} \left( \exists \mathbf{M}'\in \mathcal{W}_a : \mathscr{D}(\mathbf{M}') \leq \mathscr{D}(\mathbf{M}) \right) \text{.}
\end{equation}
Applying the union bound of probability \cite{boole1847mathematical} yields that
\begin{equation}
\begin{split} \label{eq12}
&\mathrm{Pr} \left( \exists \mathbf{M}'\in \mathcal{W}_a : \mathscr{D}(\mathbf{M}') \leq \mathscr{D}(\mathbf{M}) \right) \\
&\leq \sum_{\mathbf{M}' \in \mathcal{W}_a} \mathrm{Pr} \left( \mathscr{D}(\mathbf{M}') \leq \mathscr{D}(\mathbf{M}) \right) \text{.}
\end{split}
\end{equation}
Given that $\mathscr{D}(\mathbf{M})$ and $\mathscr{D}(\mathbf{M}')$ have been calculated in (\ref{eq8}) and (\ref{eq9}) respectively, the probability $\mathrm{Pr} \left( \mathscr{D}(\mathbf{M}') \leq \mathscr{D}(\mathbf{M}) \right)$ is calculated as follows:
%公式太长了，问题无法解决
{
\small
\begin{align} \label{eq13}
& \mathrm{Pr} \left( \mathscr{D}(\mathbf{M}') \leq \mathscr{D}(\mathbf{M}) \right) \notag \\
& = \mathrm{Pr}\bigg( \sum_{i=1}^{n/k}\sum_{j=1}^{L}{\left(y_{i,j}-h_{i,j}\operatorname{f}(x_{i,j}(\mathbf{M'}))\right)}^2 \leq \sum_{i=1}^{n/k}\sum_{j=1}^{L} n_{i,j}^2 \bigg) \notag\\
& \overset{(a)}{=} \mathrm{Pr}\bigg( \sum_{i=a}^{n/k}\sum_{j=1}^{L}{\left(y_{i,j}-h_{i,j}\operatorname{f}(x_{i,j}(\mathbf{M'}))\right)}^2 \leq \sum_{i=a}^{n/k}\sum_{j=1}^{L} n_{i,j}^2 \bigg) \notag \\
& \overset{(b)}{=} \mathrm{Pr}\bigg( \sum_{i=a}^{n/k}\sum_{j=1}^{L}{\left[h_{i,j}(\operatorname{f}(x_{i,j}(\mathbf{M}))-f(x_{i,j}(\mathbf{M'}))) + n_{i,j}\right]}^2 \notag \\
& \qquad \qquad \qquad \qquad \leq \sum_{i=a}^{n/k}\sum_{j=1}^{L} n^2_{i,j} \bigg) ,
\end{align}
}
where ($a$) establishes since $\operatorname{f}(\mathbf{x}_{i,j}(\mathbf{M}))=\operatorname{f}(\mathbf{x}_{i,j}(\mathbf{M'}))$ for $1\leq i<a$, which is proved in
Appendix \ref{property} by leveraging the property of hash function. ($b$) is obtained by applying (\ref{eq5}).
Define $U_{i,j} \triangleq \operatorname{f}(\mathbf{x}_{i,j}(\mathbf{M}))-\operatorname{f}(\mathbf{x}_{i,j}(\mathbf{M'}))$ and $V_{i,j} \triangleq h_{i,j}U_{i,j}$, and then (\ref{eq13}) can be expanded as
\begin{equation} \label{eq15}
\small
\begin{split}
&\mathrm{Pr}\bigg( \sum_{i=a}^{n/k}\sum_{j=1}^{L}{\bigg[\underbrace{h_{i,j}(\operatorname{f}(x_{i,j}(\mathbf{M}))-\operatorname{f}(x_{i,j}(\mathbf{M'})))}_{V_{i,j}} + n_{i,j}\bigg]}^2 \\
& \qquad \qquad \quad \quad \leq \sum_{i=a}^{n/k}\sum_{j=1}^{L} n^2_{i,j} \bigg) \\
&= \mathrm{Pr} \left( \sum_{i=a}^{n/k}\sum_{j=1}^{L} V^2_{i,j} + 2\sum_{i=a}^{n/k}\sum_{j=1}^{L} V_{i,j}n_{i,j} \leq 0 \right) \text{.}
\end{split}
\end{equation}
Denoting $\mathbf{V}^{L_a}$ as the random row vector composed of random variables $V_{i,j}$ with $a \leq i \leq n/k, 1 \leq j \leq L$, and $\mathbf{N}^{L_a}$ as the random row vector composed of random variables $n_{i,j}$ with $a \leq i \leq n/k, 1 \leq j \leq l_i$, (\ref{eq15}) could be rewritten into a vector form, given as
\begin{equation} \label{eq16}
\small
\begin{split}
& \mathrm{Pr} \left( \sum_{i=a}^{n/k}\sum_{j=1}^{L} V^2_{i,j} + 2\sum_{i=a}^{n/k}\sum_{j=1}^{L} V_{i,j}n_{i,j} \leq 0 \right) \\
&= \mathrm{Pr} \left( \mathbf{V}^{L_a} {\left(\mathbf{V}^{L_a} + 2{\mathbf{N}}^{L_a}\right)}^{\mathrm{T}} \leq 0 \right),
%&= \underbrace{\int_{\mathbb{R}} \cdots \int_{\mathbb{R}}}_{L_a} \mathrm{Pr} \left( \mathbf{Y}^{L_a} {\left(\mathbf{Y}^{L_a} + 2{\mathbf{N}}^{L_a}\right)}^{\mathrm{T}} \leq 0 \big| \mathbf{Y}^{L_a} \right) \\
%& \qquad\prod_{i=a}^{n/k}\prod_{j=1}^{l_i} f_{\mathcal{Y}}(V_{i,j}) \prod_{i=a}^{n/k}\prod_{j=1}^{l_i} \mathrm{d}V_{i,j} \text{.}
\end{split}
\end{equation}
which could be then expanded as
\begin{equation}\label{integral19}
\begin{aligned}
&\int_{\mathbb{R}^{L_a}}\mathrm{Pr} \left( \mathbf{V}^{L_a} {\left(\mathbf{V}^{L_a} + 2{\mathbf{N}}^{L_a}\right)}^{\mathrm{T}} \leq 0 \big| \mathbf{V}^{L_a}=\mathbf{v}^{L_a} \right)\cdot\\
&\quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \mathrm{Pr} \left( \mathbf{V}^{L_a}=\mathbf{v}^{L_a} \right) {~\mathrm{d}\mathbf{v}^{L_a}}\\
&= \int_{\mathbb{R}^{L_a}}\mathrm{Pr} \left( \mathbf{v}^{L_a} {\left(\mathbf{v}^{L_a} + 2{\mathbf{N}}^{L_a}\right)}^{\mathrm{T}} \leq 0 \right)\cdot\\
&\quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \mathrm{Pr} \left( \mathbf{V}^{L_a}=\mathbf{v}^{L_a} \right) {~\mathrm{d}\mathbf{v}^{L_a}}.
\end{aligned}
\end{equation}
With (\ref{integral19}) in hand, the next problem is to explicitly solve $\mathrm{Pr} \left( \mathbf{v}^{L_a} {\left(\mathbf{v}^{L_a} + 2{\mathbf{N}}^{L_a}\right)}^{\mathrm{T}} \leq 0 \right)$ and $\mathrm{Pr} \left( \mathbf{V}^{L_a}=\mathbf{v}^{L_a} \right)$. First, we attempt to simplify $\mathrm{Pr} \left( \mathbf{v}^{L_a} {\left(\mathbf{v}^{L_a} + 2{\mathbf{N}}^{L_a}\right)}^{\mathrm{T}} \leq 0 \right)$. By introducing two rotation matrices for $L_a$-dimensions hyperspace, we obtain Lemma \ref{lemma1} as follows:
\begin{lemma}\label{lemma1}
Given that $n_{i,j}$ is i.i.d AWGN with variance $\sigma^2$, the probability in (\ref{integral19}) can be simplified as
\begin{equation}\label{Qtransformation}
\small
\mathrm{Pr} \left( \mathbf{v}^{L_a} {\left(\mathbf{v}^{L_a} + 2{\mathbf{N}}^{L_a}\right)}^{\mathrm{T}} \leq 0 \right)=Q\left( \frac{\left\|\mathbf{v}^{L_a}\right\|}{2\sigma} \right),
\end{equation}
where $Q(\cdot)$ represents the $Q$ function, $\left\|\cdot\right\|$ represents the $\ell^2$ norm.
\end{lemma}
\begin{proof}
Please refer to Appendix \ref{rotationmatrix}.
\end{proof}
%1.0
%Aiming to make $\mathrm{Pr} \left( \mathbf{Y}^{L_a} {\left(\mathbf{Y}^{L_a} + 2{\mathbf{N}}^{L_a}\right)}^{\mathrm{T}} \leq 0 \big| \mathbf{Y}^{L_a} \right)$ easier, we introduce a rotation matrix for $L_a$-dimensions hyperspace to simplify the above expression. The rotation matrix is denoted as
%旋转矩阵的推理过程
\begin{comment}
As a rotation matrix, $\mathbf{A}$ owns the property that ${\mathbf{A}}^{\mathrm{T}} \mathbf{A} = \mathbf{I}_{L_a}$, which can be used to rewrite $\mathrm{Pr} \left( \mathbf{Y}^{L_a} {\left(\mathbf{Y}^{L_a} + 2{\mathbf{N}}^{L_a}\right)}^{\mathrm{T}} \leq 0 \big| \mathbf{Y}^{L_a} \right)$ as
\begin{equation} \label{eq19}
\begin{split}
& \mathrm{Pr} \left( \mathbf{Y}^{L_a} {\left(\mathbf{Y}^{L_a} + 2{\mathbf{N}}^{L_a}\right)}^{\mathrm{T}} \leq 0 \big| \mathbf{Y}^{L_a} \right) \\
&= \mathrm{Pr} \left( \mathbf{Y}^{L_a} \mathbf{I}_{L_a} {\left(\mathbf{Y}^{L_a} + 2{\mathbf{N}}^{L_a}\right)}^{\mathrm{T}} \leq 0 \big| \mathbf{Y}^{L_a} \right) \\
&= \mathrm{Pr} \left( \mathbf{Y}^{L_a} {\mathbf{A}}^{\mathrm{T}} \mathbf{A} {\left(\mathbf{Y}^{L_a} + 2{\mathbf{N}}^{L_a}\right)}^{\mathrm{T}} \leq 0 \big| \mathbf{Y}^{L_a} \right) \\
&= \mathrm{Pr} \left( {\left[ \mathbf{A}{\left[\mathbf{Y}^{L_a}\right]}^{\mathrm{T}} \right]}^{\mathrm{T}} \left( \mathbf{A}{\left[\mathbf{Y}^{L_a}\right]}^{\mathrm{T}} + 2\mathbf{A}{\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} \right) \leq 0 \big| \mathbf{Y}^{L_a} \right) \text{.}
\end{split}
\end{equation}
Applying (\ref{eq17}) and (\ref{eq18}) to (\ref{eq19}) results in
\begin{equation} \label{eq20}
\begin{split}
&\mathrm{Pr} \left( {\left\| \mathbf{Y}^{L_a} \right\|}^2 + 2\left\| \mathbf{Y}^{L_a} \right\| \cdot \mathbf{A}_1 {\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} \leq 0 \big| \mathbf{Y}^{L_a} \right) \\
&= \mathrm{Pr} \left( \mathbf{A}_1 {\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} \leq -\frac{\left\| \mathbf{Y}^{L_a} \right\|}{2} \bigg| \mathbf{Y}^{L_a} \right) \text{.}
\end{split}
\end{equation}
The above conditional probability can be given by the following equivalent form:
\begin{equation} \label{eq21}
\begin{split}
&\mathrm{Pr} \left( \mathbf{A}_1 {\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} \leq -\frac{\left\| \mathbf{Y}^{L_a} \right\|}{2} \bigg| \mathbf{Y}^{L_a} \right) \\
&= \overbrace{\int \cdots \int}^{L_a}_{\mathbf{A}_1 {\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} \leq -\frac{\left\| \mathbf{Y}^{L_a} \right\|}{2}} \frac{1}{{\left(\sqrt{2\pi}\sigma\right)}^{L_a}} \mathrm{e}^{-\frac{{\left\| \mathbf{N}^{L_a} \right\|}^2}{2\sigma^2}} \prod_{i=a}^{n/k} \prod_{j=1}^{l_i} \mathrm{d}n_{i,j} \text{.}
\end{split}
\end{equation}
In order to calculate the integral above easily, we introduce another rotation matrix
\begin{equation} \label{eq22}
\mathbf{B} =
\begin{bmatrix}
\mathbf{B}_1 \\
\vdots \\
\mathbf{B}_{L_a}
\end{bmatrix}
\in \mathbb{R}^{L_a \times L_a} \text{,}
\end{equation}
such that
\begin{equation} \label{eq23}
\mathbf{B}{\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} = {\bigg[ \mathbf{A}_1{\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}},\underbrace{0,\cdots,0}_{L_a-1} \bigg]}^{\mathrm{T}} \text{,}
\end{equation}
and that
\begin{equation} \label{eq24}
{\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} = \mathbf{B}^{\mathrm{T}}{\left[\mathbf{n}^{L_a}\right]}^{\mathrm{T}} \text{,}
\end{equation}
where $\mathbf{n}^{L_a}$ is the transformed vector composed of $n'_{i,j}$ with $a \leq i \leq n/k, 1 \leq j \leq l_i$.
From (\ref{eq23}) we can get that
\begin{equation} \label{eq25}
\mathbf{B}_1{\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} = \mathbf{A}_1{\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} \text{.}
\end{equation}
What's more, rotation matrix fulfiils ${\mathbf{B}}^{\mathrm{T}} \mathbf{B} = \mathbf{I}_{L_a}$. Therefore, (\ref{eq21}) can be rewritten as
\begin{equation} \label{eq26}
\begin{split}
&\overbrace{\int \cdots \int}^{L_a}_{\mathbf{A}_1 {\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} \leq -\frac{\left\| \mathbf{Y}^{L_a} \right\|}{2}} \frac{1}{{\left(\sqrt{2\pi}\sigma\right)}^{L_a}} \mathrm{e}^{-\frac{{\left\| \mathbf{N}^{L_a} \right\|}^2}{2\sigma^2}} \prod_{i=a}^{n/k} \prod_{j=1}^{l_i} \mathrm{d}n_{i,j} \\
&= \overbrace{\int \cdots \int}^{L_a}_{\mathbf{A}_1 \mathbf{B}^{\mathrm{T}} {\left[\mathbf{n}^{L_a}\right]}^{\mathrm{T}} \leq -\frac{\left\| \mathbf{Y}^{L_a} \right\|}{2}} \frac{1}{{\left(\sqrt{2\pi}\sigma\right)}^{L_a}} \mathrm{e}^{-\frac{ \mathbf{n}^{L_a}\mathbf{B}\mathbf{B}^{\mathrm{T}}{\left[\mathbf{n}^{L_a}\right]}^{\mathrm{T}} }{2\sigma^2}} \\
& \quad \; \cdot \left| \mathbf{J}\left(\mathbf{n}^{L_a}\right) \right|\prod_{i=a}^{n/k} \prod_{j=1}^{l_i} \mathrm{d}n_{i,j} \text{,}
\end{split}
\end{equation}
in which $\mathbf{J}\left(\mathbf{n}^{L_a}\right)$ is the Jacobi matrix, behaving as
\begin{equation} \label{eq27}
\mathbf{J}\left(\mathbf{n}^{L_a}\right) =
\begin{bmatrix}
\frac{\partial n'_{a,1}}{n_{a,1}} & \cdots & \frac{\partial n'_{a,1}}{n_{n/k,l_{n/k}}} \\
\vdots & \ddots & \vdots \\
\frac{\partial n'_{n/k,l_{n/k}}}{n_{a,1}} & \cdots & \frac{\partial n'_{n/k,l_n/k}}{n_{n/k,l_{n/k}}}
\end{bmatrix} \text{.}
\end{equation}
From (\ref{eq24}) we can obtain that ${\left[\mathbf{n}^{L_a}\right]}^{\mathrm{T}} = \mathbf{B}{\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}}$, therefore Jacobi matrix behaves as
\begin{equation} \label{eq28}
\mathbf{J}\left(\mathbf{n}^{L_a}\right) = \mathbf{B} \text{.}
\end{equation}
Recall that ${\mathbf{B}}^{\mathrm{T}} \mathbf{B} = \mathbf{I}_{L_a}$, we have
\begin{equation} \label{eq29}
\begin{split}
\left| \mathbf{J}\left(\mathbf{n}^{L_a}\right) \right| &= \sqrt{{\left| \mathbf{J}\left(\mathbf{n}^{L_a}\right) \right|}^2} \\
&= \sqrt{{\left|B\right|}^2} = \sqrt{\left| \mathbf{B}{\mathbf{B}}^{\mathrm{T}} \right|} = 1 \text{.}
\end{split}
\end{equation}
Together with $\mathbf{B}_1{\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} = \mathbf{A}_1{\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}}$, we have
\begin{equation} \label{eq30}
\begin{split}
\mathbf{A}_1 \mathbf{B}^{\mathrm{T}} {\left[\mathbf{n}^{L_a}\right]}^{\mathrm{T}}&=
\mathbf{B}_1 \mathbf{B}^{\mathrm{T}} {\left[\mathbf{n}^{L_a}\right]}^{\mathrm{T}}\\
&=
\begin{bmatrix}
\mathbf{B}_1 \mathbf{B}_1^{\mathrm{T}} & \mathbf{B}_1 \mathbf{B}_2^{\mathrm{T}} & \cdots & \mathbf{B}_1 \mathbf{B}_{L_a}^{\mathrm{T}}
\end{bmatrix}
{\left[\mathbf{n}^{L_a}\right]}^{\mathrm{T}} \\
&= \left[ 1, \underbrace{0,\cdots,0}_{L_a-1} \right] {\left[\mathbf{n}^{L_a}\right]}^{\mathrm{T}} \\
&= n'_{a,1} \text{.}
\end{split}
\end{equation}
Then, by substituting (\ref{eq29}) and (\ref{eq30}) into (\ref{eq26}), it is easy to verify that
\begin{equation} \label{eq31}
\small
\begin{split}
&\overbrace{\int \cdots \int}^{L_a}_{\mathbf{A}_1 \mathbf{B}^{\mathrm{T}} {\left[\mathbf{n}^{L_a}\right]}^{\mathrm{T}} \leq -\frac{\left\| \mathbf{Y}^{L_a} \right\|}{2}} \frac{1}{{\left(\sqrt{2\pi}\sigma\right)}^{L_a}} \mathrm{e}^{-\frac{ \mathbf{n}^{L_a}\mathbf{B}\mathbf{B}^{\mathrm{T}}{\left[\mathbf{n}^{L_a}\right]}^{\mathrm{T}} }{2\sigma^2}} \cdot \left| \mathbf{J}\left(\mathbf{n}^{L_a}\right) \right|\\
& \quad \; \prod_{i=a}^{n/k} \prod_{j=1}^{l_i} \mathrm{d}n'_{i,j} \\
&= \overbrace{\int \cdots \int}^{L_a}_{n'_{a,1} \leq -\frac{\left\| \mathbf{Y}^{L_a} \right\|}{2}} \frac{1}{{\left(\sqrt{2\pi}\sigma\right)}^{L_a}} \mathrm{e}^{-\frac{{\left\| \mathbf{n}^{L_a} \right\|}^2}{2\sigma^2}} \prod_{i=a}^{n/k} \prod_{j=1}^{l_i} \mathrm{d}n'_{i,j} \\
&= \int_{-\infty}^{-\frac{\left\| \mathbf{Y}^{L_a} \right\|}{2}} \frac{1}{{\left(\sqrt{2\pi}\sigma\right)}^{L_a}} \mathrm{e}^{-\frac{{\left( n'_{a,1} \right)}^2}{2\sigma^2}} \mathrm{d}n'_{a,1} = Q\left(\frac{ \left\| \mathbf{Y}^{L_a} \right\| }{2\sigma}\right) \text{.}
\end{split}
\end{equation}
\end{comment}
%条件概率的积分表达（长形式）
\begin{comment}
Applying (\ref{eq31}) to (\ref{eq15}), we have
\begin{equation} \label{eq32}
\small
\begin{split}
& \sum_{\mathbf{M'} \in \mathcal{W}_a} \underbrace{\int_{\mathbb{R}} \cdots \int_{\mathbb{R}}}_{L_a} \mathrm{Pr} \left( \mathbf{Y}^{L_a} {\left(\mathbf{Y}^{L_a} + 2{\mathbf{N}}^{L_a}\right)}^{\mathrm{T}} \leq 0 \big| \mathbf{Y}^{L_a} \right) \\
& \quad \; \prod_{i=a}^{n/k}\prod_{j=1}^{l_i} f_{\mathcal{Y}}(V_{i,j}) \prod_{i=a}^{n/k}\prod_{j=1}^{l_i} \mathrm{d}V_{i,j} \\
&= \sum_{\mathbf{M'} \in \mathcal{W}_a} \underbrace{\int_{\mathbb{R}} \cdots \int_{\mathbb{R}}}_{L_a} Q\left( \frac{\left\|\mathbf{Y}^{L_a}\right\|}{2\sigma} \right) \prod_{i=a}^{n/k}\prod_{j=1}^{l_i} f_{\mathcal{Y}}(V_{i,j}) \prod_{i=a}^{n/k}\prod_{j=1}^{l_i} \mathrm{d}V_{i,j} \text{.}
\end{split}
\end{equation}
\end{comment}
Note that \cite{Qfunction} has shown a transformation of $Q(\cdot)$, which is presented as an exponential form, given as
\begin{equation} \label{eq33}
\small
Q(x) = \frac{1}{\pi} \int_{0}^{\frac{\pi}{2}} \mathrm{exp} \left(\frac{-x^2}{2\mathrm{sin}^2\theta}\right) \mathrm{d}\theta \text{.}
\end{equation}
Adopting (\ref{eq33}) into (\ref{Qtransformation}) yields that
\begin{align} \label{eq34}
\small
% \begin{split}
& \mathrm{Pr} \left( \mathbf{v}^{L_a} {\left(\mathbf{v}^{L_a} + 2{\mathbf{N}}^{L_a}\right)}^{\mathrm{T}} \leq 0 \right) \notag\\
&= Q\left( \frac{\left\|\mathbf{v}^{L_a}\right\|}{2\sigma} \right) = \frac{1}{\pi} \int_{0}^{\frac{\pi}{2}} \mathrm{exp} \left(\frac{-{\left\|\mathbf{v}^{L_a}\right\|}^2}{8 \sigma^2 \mathrm{sin}^2\theta}\right) \mathrm{d}\theta \text{.}
% \end{split}
\end{align}
Applying (\ref{eq34}) in (\ref{integral19}) and swapping the integrates, we have
\begin{equation} \label{eq35}
\small
\begin{split}
&\mathrm{Pr} \left( \mathbf{V}^{L_a} {\left(\mathbf{V}^{L_a} + 2{\mathbf{N}}^{L_a}\right)}^{\mathrm{T}} \leq 0 \right)\\
&=\frac{1}{\pi}\int_{\mathbb{R}^{L_a}} \int_{0}^{\frac{\pi}{2}} \mathrm{exp} \left(\frac{-{\left\|\mathbf{v}^{L_a}\right\|}^2}{8 \sigma^2 \mathrm{sin}^2\theta}\right) \mathrm{d}\theta \cdot \mathrm{Pr} \left( \mathbf{V}^{L_a}=\mathbf{v}^{L_a} \right) \mathrm{d}\mathbf{v}^{L_a}\\
&= \frac{1}{\pi}\int_{0}^{\frac{\pi}{2}}\underbrace{\int_{\mathbb{R}^{L_a}} \mathrm{exp} \left(\frac{-{\left\|\mathbf{v}^{L_a}\right\|}^2}{8 \sigma^2 \mathrm{sin}^2\theta}\right) \cdot \mathrm{Pr} \left( \mathbf{V}^{L_a}=\mathbf{v}^{L_a} \right) ~\mathrm{d}\mathbf{v}^{L_a}}_{\mathcal{F}_{\text{Rayleigh}(\theta;\sigma,L_a)}} \mathrm{d}\theta\text{.}
\end{split}
\end{equation}
For the Rayleigh fading channel, we denote the multiple integrals with respect to $\mathbf{v}^{L_a}$ as $\mathcal{F}_{\text{Rayleigh}}\left(\theta;\sigma,L_a\right)$. By adopting the i.i.d of $V_{i,j}$, given as
\begin{equation}
\small
\Pr\left(\mathbf{V}^{L_a}=\mathbf{v}^{L_a}\right)=\prod_{i=a}^{n/k}\prod_{j=1}^{L}f_{V_{i,j}}\left(v_{i,j}\right),
\end{equation}
$\mathcal{F}_{\text{Rayleigh}}\left(\theta;\sigma,L_a\right)$ could be decomposed as
\begin{equation} \label{eq36}
\small
\begin{split}
& \underbrace{\int_{\mathbb{R}} \cdots \int_{\mathbb{R}}}_{L_a} \mathrm{exp} \left(\frac{-{\left\|\mathbf{v}^{L_a}\right\|}^2}{8 \sigma^2 \mathrm{sin}^2\theta}\right) \prod_{i=a}^{n/k}\prod_{j=1}^{L} f_{V_{i,j}}(v_{i,j}) \prod_{i=a}^{n/k}\prod_{j=1}^{L} \mathrm{d}v_{i,j} \\
&\overset{(a)}{=} \prod_{i=a}^{n/k}\prod_{j=1}^{L} \int_{\mathbb{R}} \mathrm{exp} \left(\frac{-v_{i,j}^2}{8 \sigma^2 \mathrm{sin}^2\theta}\right) f_{V_{i,j}}(v_{i,j})~\mathrm{d}v_{i,j} \\
&\overset{(b)}{=} { \left( \int_{\mathbb{R}} {\mathrm{exp}}{\left(-\frac{v_{a,1}^2}{{8}{\sigma^2}{\sin^2}\theta}\right)} f_{V_{a,1}}(v_{a,1}) \mathrm{d}v_{a,1} \right) }^{L_a} \text{,}
\end{split}
\end{equation}
where (a) establishes by adopting
\begin{equation}
\small
\begin{split}
\exp\left(\frac{-\|\mathbf{v}^{L_a}\|^2}{8\sigma^2 \sin^2 \theta}\right)&=\exp\left(\frac{-\sum_{i=a}^{n/k}\sum_{j=1}^{L}v^2_{i,j}}{8\sigma^2 \sin^2 \theta}\right)\\
&=\prod_{i=a}^{n/k}\prod_{j=1}^{L}\exp\left(\frac{-v^2_{i,j}}{8\sigma^2 \sin^2 \theta}\right) \text{,}
\end{split}
\end{equation}
and (b) holds for the i.i.d of the random variable $V_{i,j}$. Recall that $V_{i,j} = h_{i,j}U_{i,j}$ where $h_{i,j}$ and $U_{i,j}$ are independent with each other, the integral in (\ref{eq36}) with respect to $v_{a,1}$ could be then transformed to
\begin{equation} \label{eqreplace}
\small
\begin{split}
& \int_{\mathbb{R}} {\mathrm{exp}}{\left(-\frac{v_{a,1}^2}{{8}{\sigma^2}{\sin^2}\theta}\right)} f_{{V}_{a,1}}(v_{a,1}) \mathrm{d}v_{a,1} \\
& \xlongequal{v_{a,1} = hu} \sum_{u} p_{_U}(u) \int_{\mathbb{R}} {\mathrm{exp}}{\left(-\frac{{h^2}{u^2}}{{8}{\sigma^2}{\sin^2}\theta}\right)} g_{_1}(h)\mathrm{d}h \text{,}
\end{split}
\end{equation}
where $p_{_U}(u)$ is the probability mass function (PMF) of $U_{a,1}$ and $g_{_1}(h)$ is the probability density function (PDF) of $h$. At this point, the right-hand side (RHS) of (\ref{eqreplace}) could be explicitly calculated by the following lemma.
\begin{lemma} \label{lemma2}
For Rayleigh distribution whose PDF is $g_{_1}(h)$, the RHS of (\ref{eqreplace}) can be calculated by
\begin{equation} \label{eqq37}
\small
\sum_{i \in \Psi} \sum_{j \in \Psi} 2^{-2c} \frac{8{\sigma}^2{\sin}^2{\theta}}{{\Omega}(i-j)^2+8{\sigma}^2{\sin}^2{\theta}} \text{,}
\end{equation}
where $\Psi$ is the channel input set.
\end{lemma}
\begin{proof}
Please refer to Appendix \ref{proofray} for the detailed derivation.
\end{proof}
%上式需要解释各个概率密度的含义吗？
%Along with the defination that $V_{i,j} \triangleq h_{i,j}X_{i,j}$, we have
%带入Q函数的推导过程
\begin{comment}
\begin{equation} \label{eq34}
\tiny
\begin{split}
& \sum_{\mathbf{M'} \in \mathcal{W}_a} \underbrace{\int_{\mathbb{R}} \cdots \int_{\mathbb{R}}}_{L_a} Q\left( \frac{\left\|\mathbf{Y}^{L_a}\right\|}{2\sigma} \right) \prod_{i=a}^{n/k}\prod_{j=1}^{l_i} f_{\mathcal{Y}}(V_{i,j}) \prod_{i=a}^{n/k}\prod_{j=1}^{l_i} \mathrm{d}V_{i,j} \\
&= \sum_{\mathbf{M'} \in \mathcal{W}_a} \underbrace{\int_{\mathbb{R}} \cdots \int_{\mathbb{R}}}_{L_a} \frac{1}{\pi} \int_{0}^{\frac{\pi}{2}} \mathrm{exp} \left(\frac{-{\left\|\mathbf{Y}^{L_a}\right\|}^2}{8 \sigma^2 \mathrm{sin}^2\theta}\right) \mathrm{d}\theta \prod_{i=a}^{n/k}\prod_{j=1}^{l_i} f_{\mathcal{Y}}(V_{i,j}) \prod_{i=a}^{n/k}\prod_{j=1}^{l_i} \mathrm{d}V_{i,j} \\
&= \sum_{\mathbf{M'} \in \mathcal{W}_a} \frac{1}{\pi} \int_{0}^{\frac{\pi}{2}} \underbrace{\int_{\mathbb{R}} \cdots \int_{\mathbb{R}}}_{L_a} \mathrm{exp} \left(\frac{-{\left\|\mathbf{Y}^{L_a}\right\|}^2}{8 \sigma^2 \mathrm{sin}^2\theta}\right) \prod_{i=a}^{n/k}\prod_{j=1}^{l_i} f_{\mathcal{Y}}(V_{i,j}) \prod_{i=a}^{n/k}\prod_{j=1}^{l_i} \mathrm{d}V_{i,j} \mathrm{d}\theta \\
&= \frac{1}{\pi} \sum_{\mathbf{M'} \in \mathcal{W}_a} \int_{0}^{\frac{\pi}{2}} { \left( \int_{Y_{a,1}} {\mathrm{exp}}{\left(-\frac{Y_{a,1}^2}{{8}{\sigma^2}{\sin^2}\theta}\right)} f_{\mathcal{Y}}(Y_{a,1}) \mathrm{d}Y_{a,1} \right) }^{L_a} \mathrm{d} \theta \text{.}
\end{split}
\end{equation}
Recall that $f_{\mathcal{Y}}(V_{i,j}) = p_{_U}(U_{i,j})g(h_{i,j})$, (\ref{eq34}) becomes
\begin{equation} \label{eq35}
\tiny
\begin{split}
& \frac{1}{\pi} \sum_{\mathbf{M'} \in \mathcal{W}_a} \int_{0}^{\frac{\pi}{2}} { \left( \int_{Y_{a,1}} {\mathrm{exp}}{\left(-\frac{Y_{a,1}^2}{{8}{\sigma^2}{\sin^2}\theta}\right)} f_{\mathcal{Y}}(Y_{a,1}) \mathrm{d}Y_{a,1} \right) }^{L_a} \mathrm{d} \theta \\
&= \frac{1}{\pi} \sum_{\mathbf{M'} \in \mathcal{W}_a} \int_{0}^{\frac{\pi}{2}} \Bigg( \sum_{U_{a,1}} p_{_U}(U_{a,1}) \int_{0}^{+\infty} {\mathrm{exp}}{\left(-\frac{{h_{a,1}^2}{U_{a,1}^2}}{{8}{\sigma^2}{\sin^2}\theta}\right)} {g(h_{a,1})\mathrm{d}h_{a,1} \Bigg)}^{L_a} \mathrm{d}\theta \text{.}
\end{split}
\end{equation}
\end{comment}
As such, substituting (\ref{eqq37}) back into (\ref{eq36}) turns out that
\begin{equation} \label{eq37}
\small
\mathcal{F}_{\text{Rayleigh}}(\theta\text{;} \sigma\text{,}L_a) =
{\left( \sum_{i \in \Psi} \sum_{j \in \Psi} 2^{-2c} \frac{8{\sigma}^2{\sin}^2{\theta}}{{\Omega}(i-j)^2+8{\sigma}^2{\sin}^2{\theta}} \right)}^{L_a} \text{.}
\end{equation}
With (\ref{eq37}) in hand, the RHS of (\ref{eq12}) is transformed as
\begin{equation} \label{eq38}
\small
\begin{split}
& \sum_{\mathbf{M}' \in \mathcal{W}_a} \mathrm{Pr} \left( \mathscr{D}\left(\mathbf{M}'\right) \leq \mathscr{D}\left(\mathbf{M}\right) \right) \\
&= \frac{1}{\pi} \sum_{\mathbf{M}' \in \mathcal{W}_a} \int_{0}^{\frac{\pi}{2}} \mathcal{F}_{\text{Rayleigh}}(\theta\text{;} \sigma\text{,}L_a) \mathrm{d}\theta \text{.}
\end{split}
\end{equation}
Note that $\mathcal{F}_{\text{Rayleigh}}(\theta\text{;} \sigma\text{,}L_a)$ is inreasing with $\theta$ for $0 \leq \theta \leq \frac{\pi}{2}$ (see Appendix \ref{monotony} for the detailed proof), so we can arbitrarily choose $N+1$ values of $\theta$ such that $\theta_0=0$, $\theta_N=\frac{\pi}{2}$ and $0 < \theta_1 < \theta_2 < \cdots < \theta_{N-1} < \frac{\pi}{2}$ to explicitly upper bound the error probability. Combining these values back into (\ref{eq38}), we get the inequality as
\begin{equation} \label{eq39}
\small
\begin{split}
&\sum_{\mathbf{M'} \in \mathcal{W}_a} \mathrm{Pr} \left( \mathscr{D}\left(M'\right) \leq \mathscr{D}\left(M\right) \right) \\
&= \frac{1}{\pi} \cdot \sum_{\mathbf{M}' \in \mathcal{W}_a} \int_{0}^{\frac{\pi}{2}} \mathcal{F}_{\text{Rayleigh}}(\theta\text{;} \sigma\text{,}L_a) \mathrm{d}\theta \\
&\le \frac{\left| \mathcal{W}_a \right|}{\pi} \cdot \sum_{r=1}^{N} \int_{\theta_{r-1}}^{\theta_r} \mathcal{F}_{\text{Rayleigh}}(\theta_r\text{;} \sigma\text{,}L_a) \mathrm{d}\theta \\
&= \left| \mathcal{W}_a \right| \cdot \sum_{r=1}^{N} b_r \mathcal{F}_{\text{Rayleigh}}(\theta_r\text{;} \sigma\text{,}L_a) \text{,}
\end{split}
\end{equation}
where
\begin{equation} \label{eq40}
\small
b_r = \frac{\theta_r-\theta_{r-1}}{\pi} \text{,} \quad\left| \mathcal{W}_a \right| = (2^k-1)2^{n-ak}.
\end{equation}
Denoting
\begin{equation} \label{eq41}
\small
\mathscr{F}_{\text{Rayleigh}} \left(L_a \text{,} \sigma \right) = \sum_{r=1}^{N} b_r\mathcal{F}_{\text{Rayleigh}}(\theta_r \text{;} \sigma\text{,}L_a)\text{,}
\end{equation}
and applying (\ref{eq41}) to (\ref{eq12}) results in the explicit bound.
\end{proof}
\subsection{Bounds on the Nakagami-m Fading Channel}
\begin{theorem} \label{theorem2}
Consider Spinal codes with message length $n$, segmentation parameter $k$, modulation parameter $c$ and sufficiently large hash parameter $v$ transmitted over a flat slow Nakagami-m fading channel with mean square $\Omega$, shape parameter $m$ and AWGN variance $\sigma^2$. The average error probability given perfect CSI under ML decoding for Spinal codes can be upper bounded by
\begin{align} \label{eq42}
\small
P_e \leq 1-\prod_{a=1}^{n/k} \left(1-\epsilon_a\right) \text{,}
\end{align}
with
\begin{equation} \label{eq43}
\epsilon_a = \mathrm{min} \left\{ 1,\left(2^k-1\right)2^{n-ak} \cdot \mathscr{F}_{\text{Nakagami}}(L_a,\sigma) \right\} \text{,}
\end{equation}
\begin{equation} \label{eq44}
\mathscr{F}_{\text{Nakagami}} \left(L_a \text{,} \sigma \right) = \sum_{r=1}^{N} b_r\mathcal{F}_{\text{Nakagami}}(\theta_r \text{;} \sigma\text{,}L_a)\text{,}
\end{equation}
\begin{equation} \label{eq45}
\small
\begin{split}
& \mathcal{F}_{\text{Nakagami}}(\theta_r\text{;} \sigma\text{,}L_a) \\
&= {\left( \sum_{i \in \Psi} \sum_{j \in \Psi} 2^{-2c}
{\left(\frac{8m{\sigma}^2{\sin}^2{\theta}_r}{{\Omega}(i-j)^2+8m{\sigma}^2{\sin}^2{\theta}_r}\right)}^m \right)}^{L_a}\text{,}
\end{split}
\end{equation}
where $b_r = \frac{\theta_r-\theta_{r-1}}{\pi}$, $L_a = (n/k-a+1)L$, $L$ is the number of transmitted passes, $\theta_r$ is arbitrarily chosen with $\theta_0=0$, $\theta_N=\frac{\pi}{2}$ and $0 < \theta_1 < \theta_2 < \cdots < \theta_{N-1} < \frac{\pi}{2}$, and N represents the number of $\theta$ values which enables the adjustment of accuracy.
\end{theorem}
\begin{proof}
Please refer to Appendix \ref{proofnaka}.
\begin{comment}
The Nakagami-m distribution for channel fading $h_{i,j}$ is given by
\begin{equation} \label{eq46}
g(h_{i,j}) = \frac{2m^m}{\Gamma(m)\Omega^m} h_{i,j}^{2m-1} \text{exp} \left( \frac{-mh_{i,j}^2}{\Omega} \right) \text{,}
\end{equation}
where the Gamma function is defined as
\begin{equation} \label{eq47}
\Gamma(m) = {\int_{0}^{+\infty}} \mathrm{e}^{-t} t^{m-1} \mathrm{d}t \text{.}
\end{equation}
Substitute (\ref{eq46}) into (\ref{eq34}), denote $z = \frac{X_{a,1}^2}{8{{\sigma}^{2}}{{\sin}^{2}}\theta}$, the integral of $h_{a,1}$ is transformed as
\begin{equation}
\begin{split} \label{eq48}
& \int_{0}^{+\infty} {\mathrm{exp}}{\left(-\frac{{h_{a,1}^2}{X_{a,1}^2}}{{8}{\sigma^2}{\sin^2}\theta}\right)}g(h_{a,1})\mathrm{d}h_{a,1} \\
&= \int_{0}^{+\infty} \mathrm{e}^{-{z}{h_{a,1}^2}} \cdot \frac{2m^m}{\Gamma(m)\Omega^m} h_{a,1}^{2m-1} \mathrm{e}^ {-\frac{mh_{a,1}^2}{\Omega}} \mathrm{d}h_{a,1}\\
& \xlongequal[h_{a,1} = \sqrt{\frac{t}{z+\frac{m}{\Omega}}}]{t = \left( z+\frac{m}{\Omega} \right)h_{a,1}^2} \frac{m^m}{\Gamma(m){(z\Omega+m)}^m} {\int_{0}^{+\infty}} \mathrm{e}^{-t} t^{m-1} \mathrm{d}t \\
&= \frac{m^m}{{(z\Omega+m)}^m} \text{.}
\end{split}
\end{equation}
As is mentioned above, $z = \frac{X_{a,1}^2}{8{{\sigma}^{2}}{{\sin}^{2}}\theta}$, therefore (\ref{eq48}) comes into
\begin{equation} \label{eq49}
\begin{split}
\frac{m^m}{{(z\Omega+m)}^m} = {\left(\frac{8m{\sigma}^2{\sin}^2{\theta}}{{\Omega}X_{a.1}^2+8m{\sigma}^2{\sin}^2{\theta}}\right)}^m \text{.}
\end{split}
\end{equation}
Then the process is similar to Theorem 1, with the denotion that
\begin{equation} \label{eq50}
\begin{split}
& \mathcal{F}_{\text{Nakagami}}(\theta\text{;} \sigma\text{,}L_a) \\
&= {\left( \sum_{i=0}^{2^c-1} \sum_{j=0}^{2^c-1} 2^{-2c}
{\left(\frac{8m{\sigma}^2{\sin}^2{\theta}}{{\Omega}(i-j)^2+8m{\sigma}^2{\sin}^2{\theta}}\right)}^m \right)}^{L_a} \text{,}
\end{split}
\end{equation}
and
\begin{equation} \label{eq51}
\mathscr{F}_2 \left(L_a \text{,} \sigma \right) = \sum_{r=1}^{N} b_r\mathcal{F}_{\text{Nakagami}}(\theta_r \text{;} \sigma\text{,}L_a)\text{,}
\end{equation}
we can get the general bound.
\end{comment}
\end{proof}
%On condition that $m=1$, the expression $\mathcal{F}_{\text{Nakagami}}(\theta_r\text{;} \sigma\text{,}L_a)$ is the same as $\mathcal{F}_{\text{Rayleigh}}(\theta_r\text{;} \sigma\text{,}L_a)$, which mutually validates the conclusion.
\subsection{Bounds on the Rician Fading Channel}
\begin{theorem} \label{theorem3}
Consider Spinal codes with message length $n$, segmentation parameter $k$, modulation parameter $c$ and sufficiently large hash parameter $v$ transmitted over a flat slow Rician fading channel with mean square $\Omega$ ,shape parameter $K$ and AWGN variance $\sigma^2$. The average error probability given perfect CSI under ML decoding for Spinal codes can be upper bounded by
\begin{equation} \label{eq52}
\small
P_e \leq 1-\prod_{a=1}^{n/k} \left(1-\epsilon_a\right) \text{,}
\end{equation}
with
\begin{equation} \label{eq53}
\epsilon_a = \mathrm{min} \left\{ 1,\left(2^k-1\right)2^{n-ak} \cdot \mathscr{F}_{\text{Rician}}(L_a,\sigma) \right\} \text{,}
\end{equation}
\begin{equation} \label{eq54}
\mathscr{F}_{\text{Rician}} \left(L_a \text{,} \sigma \right) = \sum_{r=1}^{N} b_r\mathcal{F}_{\text{Rician}}(\theta_r \text{;} \sigma\text{,}L_a)\text{,}
\end{equation}
\begin{equation} \label{eq55}
\small
\begin{split}
& \mathcal{F}_{\text{Rician}}(\theta_r\text{;} \sigma\text{,}L_a) \\
&= \Bigg( \sum_{i \in \Psi} \sum_{j \in \Psi} 2^{-2c} \frac{8(K+1){\sigma}^2{\sin}^2{\theta}_r}{{\Omega}(i-j)^2+8(K+1){\sigma}^2{\sin}^2{\theta}_r} \\
& \qquad \cdot {\mathrm{exp} \left( \frac{8K(K+1){\sigma}^2{\sin}^2{\theta}_r}{{\Omega}(i-j)^2+8(K+1){\sigma}^2{\sin}^2{\theta}_r} -K\right) \Bigg)}^{L_a}\text{,}
\end{split}
\end{equation}
where $b_r = \frac{\theta_r-\theta_{r-1}}{\pi}$, $L_a = (n/k-a+1)L$, $L$ is the number of transmitted passes, $\theta_r$ is arbitrarily chosen with $\theta_0=0$, $\theta_N=\frac{\pi}{2}$ and $0 < \theta_1 < \theta_2 < \cdots < \theta_{N-1} < \frac{\pi}{2}$, and N represents the number of $\theta$ values which enables the adjustment of accuracy.
\end{theorem}
\begin{proof}
Please refer to Appendix \ref{proofrice}.
\begin{comment}
The Rician distribution for channel fading $h_{i,j}$ is given by
\begin{equation} \label{eq56}
\tiny
g(h_{i,j}) = \frac{2(K+1)h_{i,j}}{\Omega} \text{exp} \left( -K-\frac{(K+1){h_{i,j}^2}}{\Omega} \right) I_0 \left( 2\sqrt{\frac{K(K+1)}{\Omega}} h_{i,j}\right) \text{,}
\end{equation}
where the $I_0(\cdot)$ is the modified Bessel function of the first kind with order zero, given by
\begin{equation} \label{eq57}
I_0(x) = \sum_{k=0}^{+\infty} \frac{{(x^2/4)}^k}{k!\Gamma(k+1)} \text{.}
\end{equation}
Substitute (\ref{eq56}) into (\ref{eq34}) and denote $z = \frac{X_{a,1}^2}{8{{\sigma}^{2}}{{\sin}^{2}}\theta}$, the integral of $h_{a,1}$ is transformed as
\begin{equation}
\tiny
\begin{split} \label{eq58}
& \int_{0}^{+\infty} {\mathrm{exp}}{\left(-\frac{{h_{a,1}^2}{X_{a,1}^2}}{{8}{\sigma^2}{\sin^2}\theta}\right)}g(h_{a,1})\mathrm{d}h_{a,1} \\
&= \int_{0}^{+\infty} \mathrm{e}^{-{z}{h_{a,1}^2}} g(h_{a,1}) \mathrm{d}h_{a,1}\\
&= \frac{2(K+1)\mathrm{e}^{-K}}{\Omega} \int_{0}^{+\infty} h_{a,1} \mathrm{e}^{- \left( \frac{K+1}{\Omega}+a \right)h_{a,1}^2} I_0 \left( 2\sqrt{\frac{K(K+1)}{\Omega}}h_{a,1} \right) \mathrm{d}h_{a,1} \\
&= 2\mathrm{e}^{-K} \sum_{m=0}^{+\infty} \frac{ K^m{(K+1)}^{m+1} }{m!\Gamma(m+1)\Omega^{m+1}} \int_{0}^{+\infty} h_{a,1}^{2m+1} \mathrm{exp} \left( -\bigg(z+\frac{K+1}{\Omega}\bigg)h_{a,1}^2 \right) \mathrm{d}h_{a,1} \text{.}
\end{split}
\end{equation}
Let $t = \left( z+\frac{K+1}{\Omega} \right)h_{a,1}^2 \, , \, h_{a,1} = \sqrt{\frac{t}{z+\frac{K+1}{\Omega}}}$, we have
\begin{equation}
\begin{split} \label{eq59}
& \int_{0}^{+\infty} h_{a,1}^{2m+1} \mathrm{exp} \left( -\bigg(z+\frac{K+1}{\Omega}\bigg)h_{a,1}^2 \right) \mathrm{d}h_{a,1} \\
&= \int_{0}^{+\infty} \mathrm{e}^{-t} \cdot \frac{t^m}{2{\left( z + \frac{K+1}{\Omega} \right)}^{m+1}} \mathrm{d}t \\
&= \frac{\Gamma(m+1)}{2{\left( z + \frac{K+1}{\Omega} \right)}^{m+1}} \text{.}
\end{split}
\end{equation}
Equation (\ref{eq58}) can be rewritten as
\begin{equation} \label{eq60}
\begin{split}
& \int_{0}^{+\infty} {\mathrm{exp}}{\left(-\frac{{h_{a,1}^2}{X_{a,1}^2}}{{8}{\sigma^2}{\sin^2}\theta}\right)}g(h_{a,1})\mathrm{d}h_{a,1} \\
&= \frac{\mathrm{e}^{-K}}{K} \sum_{m=0}^{+\infty} \frac{1}{m!}{\left( \frac{K(K+1)}{{\Omega}z+K+1} \right)}^{m+1} \\
&= \frac{\mathrm{e}^{-K}}{K} \cdot \frac{K(K+1)}{{\Omega}z+K+1} \cdot \mathrm{exp}\left( \frac{K(K+1)}{{\Omega}z+K+1} \right) \text{.}
\end{split}
\end{equation}
Recall that $z = \frac{X_{a,1}^2}{8{{\sigma}^{2}}{{\sin}^{2}}\theta}$, (\ref{eq60}) comes into
\begin{equation}
\tiny
\begin{split}
& \frac{\mathrm{e}^{-K}}{K} \cdot \frac{K(K+1)}{{\Omega}z+K+1} \cdot \mathrm{exp}\left( \frac{K(K+1)}{{\Omega}z+K+1} \right) \\
&= \frac{8(K+1){\sigma^2}{\sin^2}\theta}{{\Omega}X_{a,1}^2+8(K+1){\sigma^2}{\sin^2}\theta} \mathrm{exp} \left( \frac{8K(K+1){\sigma^2}{\sin^2}\theta}{{\Omega}X_{a,1}^2+8(K+1){\sigma^2}{\sin^2}\theta}-K \right) \text{.}
\end{split}
\end{equation}
Then the process is similar to Theorem 1, with the denotion that
\begin{equation} \label{eq62}
\small
\begin{split}
& \mathcal{F}_{\text{Rician}}(\theta\text{;} \sigma\text{,}L_a) \\
&= \Bigg( \sum_{i=0}^{2^c-1} \sum_{j=0}^{2^c-1} 2^{-2c} \frac{8(K+1){\sigma}^2{\sin}^2{\theta}}{{\Omega}(i-j)^2+8(K+1){\sigma}^2{\sin}^2{\theta}} \\
& \qquad {\mathrm{exp} \left( \frac{8K(K+1){\sigma}^2{\sin}^2{\theta}}{{\Omega}(i-j)^2+8(K+1){\sigma}^2{\sin}^2{\theta}} -K\right) \Bigg)}^{L_a} \text{,}
\end{split}
\end{equation}
and
\begin{equation} \label{eq63}
\mathscr{F}_3 \left(L_a \text{,} \sigma \right) = \sum_{r=1}^{N} b_r\mathcal{F}_{\text{Rician}}(\theta_r \text{;} \sigma\text{,}L_a) \text{,}
\end{equation}
we can get the general bound.
\end{comment}
\end{proof}
\section{Simulation Result}\label{sectionIV}
In this section, we conduct Monte Carlo simulations to illustrate the accuracy of the upper bounds derived in Section \ref{section III}. Since the exponential nature of the ML-decoding complexity, the input message bits $n$ is selected as low as $n=8$ and the number of pass is set as $L = 6$ here for the easy ML-decoding simulation setup. The parameter $v$ is set to $v = 32$ for the implementation and experiments, demonstrated by the property \ref{property2} in Appendix \ref{property} that the hash collision will occur only once per $2^{32}$ hash function invocations on average. Furthermore, to improve the accuracy of approximation for upper bounds, we set $N=20$ and the sample size of Monte Carlo simulations to be $10^6$.
We nomalize the mean square values of all fading channels, \emph{i.e.}, $\Omega = 1$. Besides the setting of $\Omega$, for the Nakagami-m fading channel, we also set the shape parameter as $m=2$. And for Rician fading channels, the shape parameter $K$, defined as the ratio of the power contributions by line-of-sight path to the remaining multipaths, is set as $K=0.5$ and $K=1$, respectively.
\begin{figure}
\centering
\includegraphics[width= 0.95\linewidth]{total3.0out.pdf}
\caption{ Upper bounds on the error probability of Spinal codes with $n = 8, v = 32, L = 6, c = 8$ and $k = 2$ over several fading channels with $\Omega = 1$.}
\label{figure2}
\end{figure}
Fig.\ref{figure2} shows examples of the error probability for Rayleigh, Nakagami-m and Rician fading channels, respectively. All approximations are close to simulated values. We could observe that the FER curve for $K=1$ is lower than the one for $K=0.5$ in Rician fading channels, which is due to $K$ is the ratio of the energy in the specular path to the energy in the scattered paths, \emph{i.e.}, the larger $K$, the more deterministic the channel.
\section{Conclusion}\label{sectionV}
This paper analyzes the error probability of Spinal codes and derives upper bounds on the error probability under ML decoding over several fading channels, including the Rayleigh fading channel, the Nakagami-m fading channel and the Rician fading channel. Additionally, we conduct simulations for different fading channels and parameters. Our experimental examples show that the upper bounds we derived have a good performence on the estimation of the average error probablity.
The work in this paper may also inspire innovation about further research and efforts in related topics. The derived upper bounds can provide theoretical support and guidance for designing other high-efficiency coding-associated techniques, such as unequal error protection and concatenation with outer codes.
%\section{Acknowledgment}
%This work was supported in part by the National Natural Science Foundation of China under Grant 61871147 and Grant 62071141, and in part by the Shenzhen Science and Technology Program under Grant GXWD20201230155427003-20200730122528002, and in part by the Major Key Project of PCL under Grant PCL2021A03-1.
\bibliographystyle{IEEEtran}
\newpage
\bibliography{reference}
% 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]{./figures/liaimin.jpg}}]{Aimin Li}
% (Student Member, IEEE) received the B.E. degree in electronic and information engineering from Harbin Institute of Technology (Shenzhen) in 2020. He is currently
% pursuing the Ph.D. degree with the Department of Electronic Engineering, Harbin Institute of Technology (Shenzhen). His research interests include aerospace communications, advanced channel coding techniques, and wireless communications.
%\end{IEEEbiography}
%\vspace{-1cm}
%\begin{IEEEbiography}[{\includegraphics[width=1in,height=1.25in,clip,keepaspectratio]{./figures/wushaohua.jpg}}]{Shaohua Wu} (Member, IEEE) received the Ph.D. degree in communication engineering from the Harbin Institute of Technology (HIT), Harbin, China, in 2009.,From 2009 to 2011, he held a Postdoctoral position with the Department of Electronics and Information Engineering, Shenzhen Graduate School, HIT (Shenzhen), Shenzhen, China, where he has been an Associate Professor since 2012. He is also an Associate Professor of Peng Cheng Laboratory, Shenzhen. From 2014 to 2015, he was a Visiting Researcher with BBCR, University of Waterloo, Waterloo, ON, Canada. He holds more than 30 Chinese patents. He has authored or coauthored more than 100 papers in the above-mentioned areas. His current research interests include wireless image/video transmission, space communications, advanced channel coding techniques, and B5G wireless transmission technologies.
%\end{IEEEbiography}
%\vspace{-1cm}
%\begin{IEEEbiography}[{\includegraphics[width=1in,height=1.25in,clip,keepaspectratio]{./figures/jiaojian.jpg}}]{Jian Jiao} (Member, IEEE) received the M.S and Ph.D. degrees in communication engineering from the Harbin Institute of Technology (HIT), Harbin, China, in 2007 and 2011, respectively. From 2011 to 2015, he was a Postdoctoral Research Fellow of the Communication Engineering Research Centre, Shenzhen Graduate School, HIT, Shenzhen, China. From 2016 to 2017, he was a China Scholarship Council Visiting Scholar with the School of Electrical and Information Engineering, The University of Sydney, Sydney, Australia. Since 2015, he has been with the School of Electrical and Information Engineering, HIT, where he is currently an Associate Professor. His research interests include error control codes, satellite communications, and massive RA.
%\end{IEEEbiography}
%\vspace{-1cm}
%\begin{IEEEbiography}[{\includegraphics[width=1in,height=1.25in,clip,keepaspectratio]{./figures/zhangning.jpg}}]{Ning Zhang}
% (Senior Member, IEEE) received his B.S. degree from Beijing Jiaotong University, China, in 2007, his M.S. degree from Beijing University of Posts and Telecommunications in 2010, and his Ph.D. degree from the University of Waterloo, Canada, in 2015. After that, he was a postdoctoral research fellow at the University of Waterloo and the University of Toronto, Canada, respectively. He is currently an associate professor with the University of Windsor, Canada. His research interests include connected vehicles, mobile edge computing, wireless networking, and machine learning. He was a recipient of the Best Paper Awards from the \emph{IEEE Globecom} in 2014, the \emph{IEEE WCSP} in 2015, the \emph{Journal of Communications and Information Networks} in 2018, the \emph{IEEE ICC} in 2019, the \emph{IEEE Technical Committee on Transmission Access} and \emph{Optical Systems} in 2019, and the \emph{IEEE ICCC} in 2019. He serves as an Associate Editor for \emph{IEEE Internet oF Things Journal}, \emph{IEEE Transactions on Cognitive Communications and Networking}, \emph{IEEE Access}, and \emph{IET Communications}, and an Area Editor of \emph{Encyclopedia of Wireless Networks} (Springer) and \emph{Cambridge Scholars}. He also serves/served as a Guest Editor for several international journals, such as \emph{IEEE Wireless Communications}, \emph{IEEE Transactions on Cognitive Communications and Networking}, \emph{IEEE Access}, and \emph{IET Communications}. He was the Workshop Chair for MobiEdge’18 (in conjunction with IEEE WiMob 2018), CoopEdge’18 (in conjunction with \emph{IEEE EDGE 2018}), and 5G\&NTN’19 (in conjunction with \emph{IEEE EDGE 2019}).
%\end{IEEEbiography}
%
%\begin{IEEEbiography}[{\includegraphics[width=1in,height=1.25in,clip,keepaspectratio]{./figures/zhangqinyu.jpg}}]{Qinyu Zhang} (Senior Member, IEEE) received the bachelor's degree in communication engineering from the Harbin Institute of Technology (HIT), Harbin, China, in 1994 and the Ph.D. degree in biomedical and electrical engineering from the University of Tokushima, Tokushima, Japan, in 2003.,From 1999 to 2003, he was an Assistant Professor with the University of Tokushima. From 2003 to 2005, he was an Associate Professor with the Shenzhen Graduate School, HIT, and was the Founding Director of the Communication Engineering Research Center, School of Electronic and Information Engineering (EIE). Since 2005, he has been a Full Professor and the Dean of EIE School, HIT. His research interests include aerospace communications and networks, wireless communications and networks, cognitive radios, signal processing, and biomedical engineering. Dr. Zhang has been a TPC Member for the \emph{Infocom}, \emph{IEEE ICC}, \emph{IEEE Globecom}, \emph{IEEE Wireless Communications and Networking Conference}, and other flagship conferences in communications. He was an Associate Chair for \emph{Finance of the International Conference on Materials and Manufacturing Technologies 2012}, the TPC Co-Chair of the \emph{IEEE/CIC ICCC 2015}, and the Symposium Co-Chair of the \emph{CHINACOM 2011}, and the \emph{IEEE Vehicular Technology Conference 2016 (Spring)}. He was the Founding Chair of the \emph{IEEE Communications Society Shenzhen Chapter}. He is on the Editorial Board of some academic journals, such as \emph{Journal of Communication}, \emph{KSII Transactions on Internet and Information Systems}, and \emph{Science China Information Sciences}.
%\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}
\clearpage
% that's all folks
\begin{appendices}
% \setcounter{table}{0} % 将表格计数设置为0
% \setcounter{figure}{0} % 将图片计数设置为0
% \setcounter{equation}{0} % 将公式计数设置为0
% \renewcommand{\thetable}{\thesection-\arabic{table}} %在表格序号前面加上章节号，如：A-1，B-1
% \renewcommand{\theequation}{\thesection-\arabic{equation}} %在公式序号前加上章节号，如：A-1
\section{The properties of hash and related inferences}\label{property}
% \subsection{The Properties of Hash Function}
The core idea of Spinal codes is to apply a hash function sequentially to generate a pseudo-random mapping from message bits to coded bits. The hash function is expressed as
\begin{equation}
h: \{0,1\}^v \times \{0,1\}^k \rightarrow \{0,1\}^v,
\end{equation}
where $v$ and $k$ are both integers.
In Spinal codes, $h$ is chosen from a pairwise independent family of hash functions $\mathscr{H}$\cite{hashfunction} by uniformly using a random seed.
\begin{property} \label{property1}
As Perry et al. indicate \cite{2012Spinal}, the hash function employed by Spinal codes should have pairwise independent property:
\begin{equation}
\begin{split}
& \mathrm{Pr} \{ h(\mathbf{s},\mathbf{m}) = \mathbf{x},h(\mathbf{s}',\mathbf{m}') = \mathbf{x}' \} \\
&= \mathrm{Pr} \{ h(\mathbf{s},\mathbf{m}) = \mathbf{x}\} \cdot \mathrm{Pr}\{ h(\mathbf{s}',\mathbf{m}') = \mathbf{x}' \} \\
&= 2^{-2v} \text{,}
\end{split}
\end{equation}
where $(\mathbf{s},\mathbf{m})$ and $(\mathbf{s}',\mathbf{m}')$ are arbitrarily chosen, $(\mathbf{s},\mathbf{m}) \neq (\mathbf{s}',\mathbf{m}')$.
\end{property}
\begin{property} \label{property2}
A sufficiently large v will lead to a zero-approaching hash collision probability, with \cite{2012Spinal}
\begin{equation}
\begin{split}
& \mathrm{Pr} \{ h(\mathbf{s},\mathbf{m}) = h(\mathbf{s}',\mathbf{m}') \} \\
&= \sum_{\mathbf{x} \in {\{ 0,1 \}}^v} \underbrace{\mathrm{Pr} \{ h(\mathbf{s},\mathbf{m}) = \mathbf{x},h(\mathbf{s}',\mathbf{m}') = \mathbf{x}' \}}_{\mathrm{Property} \ref{property1}} \\
&= 2^v \cdot 2^{-2v} \\
&= 2^{-v} \approx 0,\ \text{iff}\ v \gg 0 \text{,}
\end{split}
\end{equation}
where $(\mathbf{s},\mathbf{m})$ and $(\mathbf{s}',\mathbf{m}')$ are arbitrarily chosen, $(\mathbf{s},\mathbf{m}) \neq (\mathbf{s}',\mathbf{m}')$.
\end{property}
Recall that $\mathbf{M}$ is the correct decoding sequence and $\mathbf{M}' \in \mathcal{W}$ is a certain wrong decoding sequence. We denote the spine values of $\mathbf{M}$ as $s(\mathbf{M}) = (\mathbf{s}_1, \mathbf{s}_2, \cdots, \mathbf{s}_{n/k})$ and the spine values of $\mathbf{M}'$ as $s(\mathbf{M}') = (\mathbf{s}'_1, \mathbf{s}'_2, \cdots, \mathbf{s}'_{n/k})$.
In (\ref{eqhash}), the process of generating $\mathbf{s}_i$ has given that $\mathbf{s}_i = h(\mathbf{s}_{i-1},\mathbf{m}_i)$, $\mathbf{s}_0=\mathbf{0}^v$. So it is natural to get that
\begin{equation}
\mathrm{Pr}( \mathbf{s}_i = \mathbf{s}'_i | \mathbf{m}_i = \mathbf{m}'_i, \mathbf{s}_{i-1} = \mathbf{s}'_{i-1} ) = 1\text{,}
\end{equation}
for $1 \leq i \leq n/k $.
Conversely, if $\mathbf{m}_i \neq \mathbf{m}'_i$ or $\mathbf{s}_{i-1} \neq \mathbf{s}'_{i-1} $ with a sufficiently large $v$, from Property \ref{property2} we can obtain that $\mathbf{s}_i \neq \mathbf{s}'_i$ for $1 \leq i \leq n/k$.
%另一种表达方式
\begin{comment}
or(don't know whether right or not)
\begin{equation}
\mathrm{Pr}( \mathbf{s}_1 \neq \mathbf{s}'_1 | \mathbf{m}_1 \neq \mathbf{m}'_1 ) = 1 \text{,}
\end{equation}
or
\begin{equation}
\mathrm{Pr}( \mathbf{s}_i \neq \mathbf{s}'_i | \mathbf{m}_i \neq \mathbf{m}'_i \text{ \ or \ } \mathbf{s}_{i-1} \neq \mathbf{s}'_{i-1} ) = 1,
\end{equation}
for $2 \leq i \leq n/k$.
\end{comment}
When $\mathbf{M}' \in \mathcal{W}_a$, $i.e.$, $\mathbf{m}_i = \mathbf{m}'_i$ for $1 \leq i < a$ and $\mathbf{m}_a \neq \mathbf{m}'_a$, we can iterate to obtain that $\mathbf{s}_i = \mathbf{s}'_i$ for $1 \leq i < a$ and $\mathbf{s}_i \neq \mathbf{s}'_i$ for $a \leq i \leq n/k$. Hence after RNGs and uniform constellation mapping, the coded symbols satisfy
\begin{equation} \label{eq(a)}
\operatorname{f}(\mathbf{x}_{i,j}(\mathbf{M})) = \operatorname{f}(\mathbf{x}_{i,j}(\mathbf{M'})) \text{,}
\end{equation}
where $1\leq i <a$.
Attributed to $\mathbf{s}_i \neq \mathbf{s}'_i$ for $a \leq i\leq n/k$, $\operatorname{f}(\mathbf{x}_{i,j}(\mathbf{M}))$ and $\operatorname{f}(\mathbf{x}_{i,j}(\mathbf{M'}))$ follow the uniform distribution over $[0,2^c-1]$, i.d.d (independent and identically distribution). %As $U_{i,j} = f(\mathbf{x}_{i,j}(\mathbf{M}))-f(\mathbf{x}_{i,j}(\mathbf{M'}))$, the probablity of $U_{i,j}$ is $p_{_\mathcal{U}}(U_{i,j}) = 2^{-2c}$.
%for $f(x_{i,j}(\mathbf{M})) \in [0,2^c-1], f(x_{i,j}(\mathbf{M}')) \in [0,2^c-1], f(x_{i,j}(\mathbf{M})),f(x_{i,j}(\mathbf{M}')) \in \mathbb{Z}$.
% By virtue of the independence of $X_{i,j}$ and $h_{i,j}$, the probability density function of $V_{i,j}$ is $f_{\mathcal{Y}}(Y_{i,j}) = p_{_U}(U_{i,j})g(h_{i,j})$.
% As $n_{i,j}$ is the AWGN, its distribution is
% \begin{equation} \label{eq14}
% f_{\mathcal{N}} \left(n_{i,j}\right) = \frac{1}{\sqrt{2\pi}\sigma} \cdot \mathrm{e}^{-\frac{n^2_{i,j}}{2\sigma^2}} \text{.}
% \end{equation}
\section{Proof of lemma \ref{lemma1}}\label{rotationmatrix}
To simplify the probability $\mathrm{Pr} \left( \mathbf{v}^{L_a} {\left(\mathbf{v}^{L_a} + 2{\mathbf{N}}^{L_a}\right)}^{\mathrm{T}} \leq 0 \right)$ in (\ref{integral19}), we introduce a rotation matrix for the $L_a$-dimensions hyperspace.
The rotation matrix is denoted as
\begin{equation} \label{eq17}
\mathbf{A} =
\begin{bmatrix}
\mathbf{A}_1 \\
\vdots \\
\mathbf{A}_{L_a}
\end{bmatrix}
\in \mathbb{R}^{L_a \times L_a} \text{,}
\end{equation}
which satisfies
\begin{equation} \label{eq18}
\mathbf{A}{\left[\mathbf{v}^{L_a}\right]}^{\mathrm{T}} = {\bigg[ \left\|\mathbf{v}^{L_a}\right\|,\underbrace{0,\cdots,0}_{L_a-1} \bigg]}^{\mathrm{T}} \text{.}
\end{equation}
As a rotation matrix, $\mathbf{A}$ owns the property that ${\mathbf{A}}^{\mathrm{T}} \mathbf{A} = \mathbf{I}_{L_a}$, which can be used to rewrite the probability as
\begin{equation} \label{eq19}
\begin{split}
& \mathrm{Pr} \left( \mathbf{v}^{L_a} {\left(\mathbf{v}^{L_a} + 2{\mathbf{N}}^{L_a}\right)}^{\mathrm{T}} \leq 0 \right) \\
&= \mathrm{Pr} \left( \mathbf{v}^{L_a} \mathbf{I}_{L_a} {\left(\mathbf{v}^{L_a} + 2{\mathbf{N}}^{L_a}\right)}^{\mathrm{T}} \leq 0 \right) \\
&= \mathrm{Pr} \left( \mathbf{v}^{L_a} {\mathbf{A}}^{\mathrm{T}} \mathbf{A} {\left(\mathbf{v}^{L_a} + 2{\mathbf{N}}^{L_a}\right)}^{\mathrm{T}} \leq 0 \right) \\
&= \mathrm{Pr} \left( {\left[ \mathbf{A}{\left[\mathbf{v}^{L_a}\right]}^{\mathrm{T}} \right]}^{\mathrm{T}} \left( \mathbf{A}{\left[\mathbf{v}^{L_a}\right]}^{\mathrm{T}} + 2\mathbf{A}{\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} \right) \leq 0 \right) \text{.}
\end{split}
\end{equation}
Applying (\ref{eq17}) and (\ref{eq18}) in (\ref{eq19}) results in
\begin{equation} \label{eq20}
\begin{split}
&\mathrm{Pr} \left( {\left\| \mathbf{v}^{L_a} \right\|}^2 + 2\left\| \mathbf{v}^{L_a} \right\| \cdot \mathbf{A}_1 {\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} \leq 0 \right) \\
&= \mathrm{Pr} \left( \mathbf{A}_1 {\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} \leq -\frac{\left\| \mathbf{v}^{L_a} \right\|}{2} \right) \text{,}
\end{split}
\end{equation}
the RHS of which is given by the following equivalent form:
\begin{equation} \label{eq21}
\begin{split}
&\mathrm{Pr} \left( \mathbf{A}_1 {\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} \leq -\frac{\left\| \mathbf{v}^{L_a} \right\|}{2} \right) \\
&= \overbrace{\int \cdots \int}^{L_a}_{_{\mathbf{A}_1 {\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} \leq -\frac{\left\| \mathbf{v}^{L_a} \right\|}{2}}} \frac{1}{{\left(\sqrt{2\pi}\sigma\right)}^{L_a}} \mathrm{e}^{-\frac{{\left\| \mathbf{N}^{L_a} \right\|}^2}{2\sigma^2}} \prod_{i=a}^{n/k} \prod_{j=1}^{L} \mathrm{d}n_{i,j} \text{.}
\end{split}
\end{equation}
In order to further simplify the $L_a$-fold integral above, we introduce another rotation matrix $\mathbf{B}$, which is denoted as
\begin{equation} \label{eq22}
\mathbf{B} =
\begin{bmatrix}
\mathbf{B}_1 \\
\vdots \\
\mathbf{B}_{L_a}
\end{bmatrix}
\in \mathbb{R}^{L_a \times L_a} \text{,}
\end{equation}
such that
\begin{equation} \label{eq23}
\mathbf{B}{\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} = {\bigg[ \mathbf{A}_1{\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}},\underbrace{0,\cdots,0}_{L_a-1} \bigg]}^{\mathrm{T}} \text{,}
\end{equation}
and that
\begin{equation} \label{eq24}
{\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} = \mathbf{B}^{\mathrm{T}}{\left[\mathbf{n}^{L_a}\right]}^{\mathrm{T}} \text{,}
\end{equation}
where $\mathbf{n}^{L_a}$ is the transformed vector composed of $n'_{i,j}$ with $a \leq i \leq n/k, 1 \leq j \leq L$.
From (\ref{eq23}) we can get that
\begin{equation} \label{eq25}
\mathbf{B}_1{\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} = \mathbf{A}_1 {\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} \text{.}
\end{equation}
Substitute (\ref{eq24}) into (\ref{eq25}), we have that
\begin{align} \label{eq30}
% \begin{split}
&\mathbf{A}_1 {\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} = \mathbf{B}_1{\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} \notag\\
&=
\mathbf{B}_1 \mathbf{B}^{\mathrm{T}} {\left[\mathbf{n}^{L_a}\right]}^{\mathrm{T}} \notag\\
&=
\begin{bmatrix}
\mathbf{B}_1 \mathbf{B}_1^{\mathrm{T}} & \mathbf{B}_1 \mathbf{B}_2^{\mathrm{T}} & \cdots & \mathbf{B}_1 \mathbf{B}_{L_a}^{\mathrm{T}}
\end{bmatrix}
{\left[\mathbf{n}^{L_a}\right]}^{\mathrm{T}} \notag\\
& \overset{(c)}{=} \left[ 1, \underbrace{0,\cdots,0}_{L_a-1} \right] {\left[\mathbf{n}^{L_a}\right]}^{\mathrm{T}} = n'_{a,1} \text{,}
% \end{split}
\end{align}
where (c) is obtained from the property that rotation matrix fulfills $\mathbf{B}{\mathbf{B}}^{\mathrm{T}} = \mathbf{I}_{L_a}$.
From (\ref{eq24}) we can also derive that
\begin{equation} \label{eqn}
\begin{split}
{\left\| \mathbf{N}^{L_a} \right\|}^2 &= {\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}}\mathbf{N}^{L_a}\\
&= \mathbf{n}^{L_a}\mathbf{B}\mathbf{B}^{\mathrm{T}}{\left[\mathbf{n}^{L_a}\right]}^{\mathrm{T}} \overset{(c)}{=} {\left\| \mathbf{n}^{L_a} \right\|}^2 \text{,}
\end{split}
\end{equation}
where (c) establishes by adopting $\mathbf{B} {\mathbf{B}}^{\mathrm{T}} = \mathbf{I}_{L_a}$.
Together with (\ref{eq30}) and (\ref{eqn}), (\ref{eq21}) can be rewritten as \cite{Jacobian}
\begin{equation} \label{eq26}
\small
\begin{split}
&\overbrace{\int \cdots \int}^{L_a}_{\mathbf{A}_1 {\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} \leq -\frac{\left\| \mathbf{v}^{L_a} \right\|}{2}} \frac{1}{{\left(\sqrt{2\pi}\sigma\right)}^{L_a}} \mathrm{e}^{-\frac{{\left\| \mathbf{N}^{L_a} \right\|}^2}{2\sigma^2}} \prod_{i=a}^{n/k} \prod_{j=1}^{L} \mathrm{d}n_{i,j} \\
&= \overbrace{\int \cdots \int}^{L_a}_{n'_{a,1} \leq -\frac{\left\| \mathbf{v}^{L_a} \right\|}{2}} \frac{1}{{\left(\sqrt{2\pi}\sigma\right)}^{L_a}} \mathrm{e}^{-\frac{ {\left\| \mathbf{n}^{L_a} \right\|}^2 }{2\sigma^2}} \left| \mathbf{J}\left(\mathbf{n}^{L_a}\right) \right|\prod_{i=a}^{n/k} \prod_{j=1}^{L} \mathrm{d}n'_{i,j} \text{,}
\end{split}
\end{equation}
in which $\mathbf{J}\left(\mathbf{n}^{L_a}\right)$ is the Jacobi matrix, behaving as
\begin{equation} \label{eq27}
\mathbf{J}\left(\mathbf{n}^{L_a}\right) =
\begin{bmatrix}
\frac{\partial n'_{a,1}}{n_{a,1}} & \cdots & \frac{\partial n'_{a,1}}{n_{n/k,L}} \\
\vdots & \ddots & \vdots \\
\frac{\partial n'_{n/k,L}}{n_{a,1}} & \cdots & \frac{\partial n'_{n/k,L}}{n_{n/k,L}}
\end{bmatrix} \text{.}
\end{equation}
Since ${\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}} = \mathbf{B}^{\mathrm{T}}{\left[\mathbf{n}^{L_a}\right]}^{\mathrm{T}}$, we get ${\left[\mathbf{n}^{L_a}\right]}^{\mathrm{T}} = \mathbf{B}{\left[\mathbf{N}^{L_a}\right]}^{\mathrm{T}}$, therefore the Jacobi matrix behaves as
\begin{equation} \label{eq28}
\mathbf{J}\left(\mathbf{n}^{L_a}\right) = \mathbf{B} \text{.}
\end{equation}
Recalling that $\mathbf{B} {\mathbf{B}}^{\mathrm{T}} = \mathbf{I}_{L_a}$, we have
\begin{equation} \label{eq29}
\begin{split}
\left| \mathbf{J}\left(\mathbf{n}^{L_a}\right) \right| &= \sqrt{{\left| \mathbf{J}\left(\mathbf{n}^{L_a}\right) \right|}^2} \\
&= \sqrt{{\left|\mathbf{B}\right|}^2} = \sqrt{\left| \mathbf{B}{\mathbf{B}}^{\mathrm{T}} \right|} = 1 \text{.}
\end{split}
\end{equation}
Then, by substituting (\ref{eq29}) into (\ref{eq26}), we could verify that
\begin{equation} \label{eq31}
\small
\begin{split}
&\overbrace{\int \cdots \int}^{L_a}_{n'_{a,1} \leq -\frac{\left\| \mathbf{v}^{L_a} \right\|}{2}} \frac{1}{{\left(\sqrt{2\pi}\sigma\right)}^{L_a}} \mathrm{e}^{-\frac{ {\left\| \mathbf{n}^{L_a} \right\|}^2 }{2\sigma^2}} \left| \mathbf{J}\left(\mathbf{n}^{L_a}\right) \right|\prod_{i=a}^{n/k} \prod_{j=1}^{L} \mathrm{d}n'_{i,j} \\
&= \overbrace{\int \cdots \int}^{L_a}_{n'_{a,1} \leq -\frac{\left\| \mathbf{v}^{L_a} \right\|}{2}} \frac{1}{{\left(\sqrt{2\pi}\sigma\right)}^{L_a}} \mathrm{e}^{-\frac{{\left\| \mathbf{n}^{L_a} \right\|}^2}{2\sigma^2}} \prod_{i=a}^{n/k} \prod_{j=1}^{L} \mathrm{d}n'_{i,j} \\
&= \int_{-\infty}^{-\frac{\left\| \mathbf{v}^{L_a} \right\|}{2}} \frac{1}{{\sqrt{2\pi}\sigma}} \mathrm{e}^{-\frac{{\left( n'_{a,1} \right)}^2}{2\sigma^2}} \mathrm{d}n'_{a,1} = Q\left(\frac{ \left\| \mathbf{v}^{L_a} \right\| }{2\sigma}\right) \text{.}
\end{split}
\end{equation}
\section{Proof of Lemma \ref{lemma2}} \label{proofray}
The central problem is to solve out the integral of $h$ in the RHS of (\ref{eqreplace}).
Recall that $h$ follows Rayleigh distribution, whose PDF is given as
\begin{equation} \label{eq6}
g_{_1}\left(h\right) = \frac{2h}{\Omega} \mathrm{exp} \left( \frac{-h^2}{\Omega} \right) \text{.}
\end{equation}
Insert (\ref{eq6}) into the RHS of (\ref{eqreplace}), the integral is transformed as
\begin{equation}
\begin{split}
& \int_{h} {\mathrm{exp}}{\left(-\frac{{h^2}{u^2}}{{8}{\sigma^2}{\sin^2}\theta}\right)} g_{_1}(h)\mathrm{d}h \\
&= \int_{0}^{+\infty} {\mathrm{exp}}{\left(-\frac{{h^2}{u^2}}{{8}{\sigma^2}{\sin^2}\theta}\right)} \cdot \frac{2h}{\Omega} \mathrm{exp} \left( \frac{-h^2}{\Omega} \right)\mathrm{d}h \\
& \xlongequal{z = \frac{u^2}{8{{\sigma}^{2}}{{\sin}^{2}}\theta}} \frac{1}{\Omega} \int_{0}^{+\infty} \mathrm{exp}\left[- \left( z+\frac{1}{\Omega} \right) {h^2}\right] \mathrm{d}{h^2} \\
&=\frac{1}{z\Omega+1} =\frac{8{\sigma^2}{\sin^2}\theta}{{\Omega}u^2+8{\sigma^2}{\sin^2}\theta} \text{,}
\end{split}
\end{equation}
with which the RHS of (\ref{eqreplace}) becomes
\begin{equation} \label{eq64}
\begin{split}
& \sum_{u} p_{_U}(u) \int_{h} {\mathrm{exp}}{\left(-\frac{{h^2}{u^2}}{{8}{\sigma^2}{\sin^2}\theta}\right)} g_1(h)\mathrm{d}h\\
&= \sum_{u} p_{_U}(u) \cdot \frac{8{\sigma^2}{\sin^2}\theta}{{\Omega}u^2+8{\sigma^2}{\sin^2}\theta} \text{.}
\end{split}
\end{equation}
Recall that $U = U_{a,1} = \operatorname{f}(\mathbf{x}_{a,1}(\mathbf{M})) - \operatorname{f}(\mathbf{x}_{a,1}(\mathbf{M'}))$. Appendix \ref{property} draws the conclusion that both $\operatorname{f}(\mathbf{x}_{a,1}(\mathbf{M}))$ and $\operatorname{f}(\mathbf{x}_{a,1}(\mathbf{M'}))$ follow the uniform distribution over $[0,2^c-1]$, i.d.d. As a result, (\ref{eq64}) extends to
\begin{equation}
\begin{split}
&\sum_{u} p_{_U}(u) \cdot \frac{8{\sigma^2}{\sin^2}\theta}{{\Omega}u^2+8{\sigma^2}{\sin^2}\theta} \\
&\xlongequal[Y = \operatorname{f}(\mathbf{x}_{a,1}(\mathbf{M'}))]{X = \operatorname{f}(\mathbf{x}_{a,1}(\mathbf{M}))} \sum_{i \in \Psi} \sum_{j \in \Psi} p_{_X}(i)p_{_Y}(j) \frac{8{\sigma}^2{\sin}^2{\theta}}{{\Omega}(i-j)^2+8{\sigma}^2{\sin}^2{\theta}} \\
&= \sum_{i \in \Psi} \sum_{j \in \Psi} 2^{-2c} \frac{8{\sigma}^2{\sin}^2{\theta}}{{\Omega}(i-j)^2+8{\sigma}^2{\sin}^2{\theta}} \text{.}
\end{split}
\end{equation}
\section{Proof Of Monotony} \label{monotony}
\begin{equation} \label{eq66}
\small
\mathcal{F}_{\text{Rayleigh}}(\theta\text{;} \sigma\text{,}L_a) =
{\left( \sum_{i \in \Psi} \sum_{j \in \Psi} 2^{-2c} \frac{8{\sigma}^2{\sin}^2{\theta}}{{\Omega}(i-j)^2+8{\sigma}^2{\sin}^2{\theta}} \right)}^{L_a} \text{.}
\end{equation}
The monotony of $\mathcal{F}_{\text{Rayleigh}}(\theta\text{;} \sigma\text{,}L_a)$ equals to the monotony of
\begin{equation}
F_{\text{Rayleigh}}(\theta) \triangleq \frac{8{\sigma}^2{\sin}^2{\theta}}{{\Omega}(i-j)^2+8{\sigma}^2{\sin}^2{\theta}}\text{,}
\end{equation}
where $\theta \in [0,\frac{\pi}{2}]$.
Denoting $8{\sigma}^2 = b$ and ${\Omega}(i-j)^2 = k$, the derivative of $F(\theta)$ is given by
\begin{equation} \label{derivation}
\frac{\mathrm{d} F_{\text{Rayleigh}}(\theta)}{\mathrm{d} \theta} = 2kb \frac{\mathrm{sin}\theta \mathrm{cos}\theta}{{(k+\mathrm{sin}^2\theta)}^2} \geq 0 \text{,}
\end{equation}
which indicates the monotonically increasing property of function $F_{\text{Rayleigh}}(\theta)$.
As a consequence, $\mathcal{F}_{\text{Rayleigh}}(\theta\text{;} \sigma\text{,}L_a)$ is a monotonically increasing function.
\begin{equation} \label{eqq50}
\begin{split}
& \mathcal{F}_{\text{Nakagami}}(\theta\text{;} \sigma\text{,}L_a) \\
&= {\left( \sum_{i \in \Psi} \sum_{j \in \Psi} 2^{-2c}
{\left(\frac{8m{\sigma}^2{\sin}^2{\theta}}{{\Omega}(i-j)^2+8m{\sigma}^2{\sin}^2{\theta}}\right)}^m \right)}^{L_a} \text{.}
\end{split}
\end{equation}
Because $m \geq 0.5$, the monotony of $\mathcal{F}_{\text{Nakagami}}(\theta\text{;} \sigma\text{,}L_a)$ equals to the monotony of
\begin{equation}
F_{\text{Nakagami}}(\theta) \triangleq \frac{8m{\sigma}^2{\sin}^2{\theta}}{{\Omega}(i-j)^2+8m{\sigma}^2{\sin}^2{\theta}} \textbf{,}
\end{equation}
where $\theta \in [0,\frac{\pi}{2}]$.
Similar to the proof for $\mathcal{F}_{\text{Rayleigh}}(\theta\text{;} \sigma\text{,}L_a)$, we denote that $8m{\sigma}^2 = b$ and ${\Omega}(i-j)^2 = k$. Since the result is the same as (\ref{derivation}), $\mathcal{F}_{\text{Nakagami}}(\theta\text{;} \sigma\text{,}L_a)$ is also a monotonically increasing function.
\begin{equation} \label{eqq62}
\small
\begin{split}
& \mathcal{F}_{\text{Rician}}(\theta\text{;} \sigma\text{,}L_a) \\
&= \Bigg( \sum_{i \in \Psi} \sum_{j \in \Psi} 2^{-2c} \frac{8(K+1){\sigma}^2{\sin}^2{\theta}}{{\Omega}(i-j)^2+8(K+1){\sigma}^2{\sin}^2{\theta}} \\
& \qquad \cdot {\mathrm{exp} \left( \frac{8K(K+1){\sigma}^2{\sin}^2{\theta}}{{\Omega}(i-j)^2+8(K+1){\sigma}^2{\sin}^2{\theta}} -K\right) \Bigg)}^{L_a} \text{,}
\end{split}
\end{equation}
Because exponential function is monotonically increasing, the monotony of $\mathcal{F}_{\text{Rician}}(\theta\text{;} \sigma\text{,}L_a)$ equals to the monotony of
\begin{equation}
F_{\text{Rician}}(\theta) \triangleq \frac{8(K+1){\sigma}^2{\sin}^2{\theta}}{{\Omega}(i-j)^2+8(K+1){\sigma}^2{\sin}^2{\theta}}\text{.}
\end{equation}
where $\theta \in [0,\frac{\pi}{2}]$.
Similar to the proof for $\mathcal{F}_{\text{Rayleigh}}(\theta\text{;} \sigma\text{,}L_a)$, we denote $8(K+1){\sigma}^2 = b$ and ${\Omega}(i-j)^2 = k$. As the result is the same as $(\ref{derivation}), \mathcal{F}_{\text{Rician}}(\theta\text{;} \sigma\text{,}L_a)$ is also a monotonically increasing function.
\section{Proof of Theorem \ref{theorem2}} \label{proofnaka}
% \setcounter{table}{0} % 将表格计数设置为0
% \setcounter{figure}{0} % 将图片计数设置为0
% \setcounter{equation}{0} % 将公式计数设置为0
%顺序编号
For Spinal codes over the Nakagami-m fading channel, the only difference is to replace Rayleigh distribution with Nakagami-m distribution. The PDF Nakagami-m distribution is given by
\begin{equation} \label{eq46}
g_{_2}(h) = \frac{2m^m}{\Gamma(m)\Omega^m} h^{2m-1} \text{exp} \left( \frac{-mh^2}{\Omega} \right) \text{,}
\end{equation}
where $h$ is the channel fading parameter, $m \geq 0.5$ and the Gamma function is defined as
\begin{equation} \label{eq47}
\Gamma(m) = {\int_{0}^{+\infty}} \mathrm{e}^{-t} t^{m-1} \mathrm{d}t \text{.}
\end{equation}
Substitute (\ref{eq46}) into the integral of (\ref{eqreplace}) and denote $z = \frac{u^2}{8{{\sigma}^{2}}{{\sin}^{2}}\theta}$, after which the integral of $h$ is transformed as
\begin{equation}
\begin{split} \label{eq48}
& \int_{h} {\mathrm{exp}}{\left(-\frac{{h^2}{u^2}}{{8}{\sigma^2}{\sin^2}\theta}\right)} g_{_2}(h)\mathrm{d}h \\
&= \int_{0}^{+\infty} \mathrm{e}^{-{z}{h^2}} \cdot \frac{2m^m}{\Gamma(m)\Omega^m} h^{2m-1} \mathrm{e}^ {-\frac{mh^2}{\Omega}} \mathrm{d}h\\
& \xlongequal[h = \sqrt{\frac{t}{z+\frac{m}{\Omega}}}]{t = \left( z+\frac{m}{\Omega} \right)h^2} \frac{m^m}{\Gamma(m){(z\Omega+m)}^m} {\int_{0}^{+\infty}} \mathrm{e}^{-t} t^{m-1} \mathrm{d}t \\
&= \frac{m^m}{{(z\Omega+m)}^m} \text{.}
\end{split}
\end{equation}
As mentioned above, $z = \frac{u^2}{8{{\sigma}^{2}}{{\sin}^{2}}\theta}$, substitute $z$ into (\ref{eq48}):
\begin{equation} \label{eq49}
\begin{split}
\frac{m^m}{{(z\Omega+m)}^m} = {\left(\frac{8m{\sigma}^2{\sin}^2{\theta}}{{\Omega}u^2+8m{\sigma}^2{\sin}^2{\theta}}\right)}^m \text{.}
\end{split}
\end{equation}
Then the process is similar to Theorem \ref{theorem1}, by denoting that
\begin{equation} \label{eq50}
\begin{split}
& \mathcal{F}_{\text{Nakagami}}(\theta\text{;} \sigma\text{,}L_a) \\
&= {\left( \sum_{i \in \Psi} \sum_{j \in \Psi} 2^{-2c}
{\left(\frac{8m{\sigma}^2{\sin}^2{\theta}}{{\Omega}(i-j)^2+8m{\sigma}^2{\sin}^2{\theta}}\right)}^m \right)}^{L_a} \text{,}
\end{split}
\end{equation}
which is monotonically increasing proved in Appendix \ref{monotony}, and that
\begin{equation} \label{eq51}
\mathscr{F}_{\text{Nakagami}} \left(L_a \text{,} \sigma \right) = \sum_{r=1}^{N} b_r\mathcal{F}_{\text{Nakagami}}(\theta_r \text{;} \sigma\text{,}L_a)\text{,}
\end{equation}
we can get the explicit bound.
%A-1编号
\begin{comment}
For Spinal codes over Nakagami-m fading channel, the only difference is to replace the Rayleigh distribution with Nakagami-m distribution.The Nakagami-m distribution for channel fading $h_{i,j}$ is given by
\begin{equation} \label{A-1}
g(h_{i,j}) = \frac{2m^m}{\Gamma(m)\Omega^m} h_{i,j}^{2m-1} \text{exp} \left( \frac{-mh_{i,j}^2}{\Omega} \right) \text{,}
\end{equation}
where the Gamma function is defined as
\begin{equation} \label{A-2}
\Gamma(m) = {\int_{0}^{+\infty}} \mathrm{e}^{-t} t^{m-1} \mathrm{d}t \text{.}
\end{equation}
Substitute (\ref{A-1}) into (\ref{eq34}), denote $z = \frac{X_{a,1}^2}{8{{\sigma}^{2}}{{\sin}^{2}}\theta}$, the integral of $h_{a,1}$ is transformed as
\begin{equation} \label{A-3}
\begin{split}
& \int_{0}^{+\infty} {\mathrm{exp}}{\left(-\frac{{h_{a,1}^2}{X_{a,1}^2}}{{8}{\sigma^2}{\sin^2}\theta}\right)}g(h_{a,1})\mathrm{d}h_{a,1} \\
&= \int_{0}^{+\infty} \mathrm{e}^{-{z}{h_{a,1}^2}} \cdot \frac{2m^m}{\Gamma(m)\Omega^m} h_{a,1}^{2m-1} \mathrm{e}^ {-\frac{mh_{a,1}^2}{\Omega}} \mathrm{d}h_{a,1}\\
& \xlongequal[h_{a,1} = \sqrt{\frac{t}{z+\frac{m}{\Omega}}}]{t = \left( z+\frac{m}{\Omega} \right)h_{a,1}^2} \frac{m^m}{\Gamma(m){(z\Omega+m)}^m} {\int_{0}^{+\infty}} \mathrm{e}^{-t} t^{m-1} \mathrm{d}t \\
&= \frac{m^m}{{(z\Omega+m)}^m} \text{.}
\end{split}
\end{equation}
As is mentioned above, $z = \frac{X_{a,1}^2}{8{{\sigma}^{2}}{{\sin}^{2}}\theta}$, therefore (\ref{A-3}) comes into
\begin{equation}
\begin{split}
\frac{m^m}{{(z\Omega+m)}^m} = {\left(\frac{8m{\sigma}^2{\sin}^2{\theta}}{{\Omega}X_{a.1}^2+8m{\sigma}^2{\sin}^2{\theta}}\right)}^m \text{.}
\end{split}
\end{equation}
Then the process is similar to Theorem 1, with the denotion that
\begin{equation}
\begin{split}
& \mathcal{F}_{\text{Nakagami}}(\theta\text{;} \sigma\text{,}L_a) \\
&= {\left( \sum_{i=0}^{2^c-1} \sum_{j=0}^{2^c-1} 2^{-2c}
{\left(\frac{8m{\sigma}^2{\sin}^2{\theta}}{{\Omega}(i-j)^2+8m{\sigma}^2{\sin}^2{\theta}}\right)}^m \right)}^{L_a} \text{,}
\end{split}
\end{equation}
and
\begin{equation}
\mathscr{F}_2 \left(L_a \text{,} \sigma \right) = \sum_{r=1}^{N} b_r\mathcal{F}_{\text{Nakagami}}(\theta_r \text{;} \sigma\text{,}L_a)\text{,}
\end{equation}
we can get the general bound.
\end{comment}
\section{Proof of Theorem \ref{theorem3}} \label{proofrice}
% \setcounter{equation}{0} % 将公式计数设置为0
%顺序编号
For Spinal codes over the Rician fading channel, the only difference is to replace Rayleigh distribution with Rician distribution. The PDF of Rician distribution is given by
\begin{equation} \label{eq56}
\small
g_{_3}(h) = \frac{2(K+1)h}{{\Omega}\, {\mathrm{exp} \left( K+\frac{(K+1){h^2}}{\Omega} \right)}} I_0 \left( 2\sqrt{\frac{K(K+1)}{\Omega}} h\right) \text{,}
\end{equation}
where $h$ is the channel fading parameter, and $I_0(\cdot)$ is the modified Bessel function of the first kind with order zero, given by
\begin{equation} \label{eq57}
I_0(x) = \sum_{k=0}^{+\infty} \frac{{(x^2/4)}^k}{k!\Gamma(k+1)} \text{.}
\end{equation}
%长公式的格式
\begin{comment}
The Rician distribution for channel fading $h_{i,j}$ is given by (\ref{eq56}), shown at the bottom of the next page,
\begin{figure*}
\centering
\begin{equation} \label{eq56}
g(h_{i,j}) = \frac{2(K+1)h_{i,j}}{\Omega \cdot \mathrm{exp} \left( K+\frac{(K+1){h_{i,j}^2}}{\Omega} \right)} I_0 \left( 2\sqrt{\frac{K(K+1)}{\Omega}} h_{i,j}\right) \text{.}
\end{equation}
{\noindent} \rule[-10pt]{18cm}{0.05em}
\end{figure*}
\end{comment}
Substitute (\ref{eq56}) into (\ref{eqreplace}) and denote $z = \frac{u^2}{8{{\sigma}^{2}}{{\sin}^{2}}\theta}$, then the integral of $h$ is transformed as
\begin{equation}
\small
\begin{split} \label{eq58}
& \int_{h} {\mathrm{exp}}{\left(-\frac{{h^2}{u^2}}{{8}{\sigma^2}{\sin^2}\theta}\right)} g_{_3}(h)\mathrm{d}h = \int_{0}^{+\infty} \mathrm{e}^{-{z}{h^2}} g_{_3}(h) \mathrm{d}h\\
&= \frac{2(K+1)}{{\Omega} \mathrm{e}^{K}} {\int_{0}^{+\infty}} {h\mathrm{e}^{- \left( \frac{K+1}{\Omega}+z \right)h^2} I_0 \left( 2\sqrt{\frac{K(K+1)}{\Omega}}h \right) \mathrm{d}h} \\
&= \frac{2}{\mathrm{e}^{K}} \sum_{m=0}^{+\infty} \frac{ K^m{(K+1)}^{m+1} }{m!\Gamma(m+1)\Omega^{m+1}} {\int_{0}^{+\infty}} h^{2m+1} \mathrm{e}^{- \left( \frac{K+1}{\Omega}+z \right)h^2} \mathrm{d}h \text{.}
\end{split}
\end{equation}
%\mathrm{exp} \left( -\bigg(z+\frac{K+1}{\Omega}\bigg)h^2 \right)
Let $t = \left( z+\frac{K+1}{\Omega} \right)h^2 $, $ h = \sqrt{\frac{t}{z+\frac{K+1}{\Omega}}}$, we have
\begin{equation}
\begin{split} \label{eq59}
& {\int_{0}^{+\infty}} h^{2m+1} \mathrm{e}^{- \left( \frac{K+1}{\Omega}+z \right)h^2} \mathrm{d}h \\
&= \int_{0}^{+\infty} \mathrm{e}^{-t} \cdot \frac{t^m}{2{\left( z + \frac{K+1}{\Omega} \right)}^{m+1}} \mathrm{d}t \\
&= \frac{\Gamma(m+1)}{2{\left( z + \frac{K+1}{\Omega} \right)}^{m+1}} \text{.}
\end{split}
\end{equation}
With (\ref{eq59}) in hand, (\ref{eq58}) can be rewritten as
\begin{equation} \label{eq60}
\begin{split}
& \int_{h} {\mathrm{exp}}{\left(-\frac{{h^2}{u^2}}{{8}{\sigma^2}{\sin^2}\theta}\right)} g_{_3}(h)\mathrm{d}h \\
&= \frac{1}{K\mathrm{e}^{K}} \cdot \frac{K(K+1)}{{\Omega}z+K+1} \cdot \sum_{m=0}^{+\infty} \frac{1}{m!}{\left( \frac{K(K+1)}{{\Omega}z+K+1} \right)}^{m} \\
&= \frac{1}{K\mathrm{e}^{K}} \cdot \frac{K(K+1)}{{\Omega}z+K+1} \cdot \mathrm{exp}\left( \frac{K(K+1)}{{\Omega}z+K+1} \right) \text{.}
\end{split}
\end{equation}
Now that $z = \frac{u^2}{8{{\sigma}^{2}}{{\sin}^{2}}\theta}$, (\ref{eq60}) comes into
\begin{equation}
\small
\begin{split}
& \frac{1}{K\mathrm{e}^{K}} \cdot \frac{K(K+1)}{{\Omega}z+K+1} \cdot \mathrm{exp}\left( \frac{K(K+1)}{{\Omega}z+K+1} \right) \\
&= \frac{8\mathrm{e}^{-K}(K+1){\sigma^2}{\sin^2}\theta}{{\Omega}u^2+8(K+1){\sigma^2}{\sin^2}\theta} \mathrm{exp} \left( \frac{8K(K+1){\sigma^2}{\sin^2}\theta}{{\Omega}u^2+8(K+1){\sigma^2}{\sin^2}\theta} \right) \text{.}
\end{split}
\end{equation}
Then the process is similar to Theorem \ref{theorem1}. By denoting that
\begin{equation} \label{eq62}
\small
\begin{split}
& \mathcal{F}_{\text{Rician}}(\theta\text{;} \sigma\text{,}L_a) \\
&= \Bigg( \sum_{i \in \Psi} \sum_{j \in \Psi} 2^{-2c} \frac{8(K+1){\sigma}^2{\sin}^2{\theta}}{{\Omega}(i-j)^2+8(K+1){\sigma}^2{\sin}^2{\theta}} \\
& \qquad \cdot {\mathrm{exp} \left( \frac{8K(K+1){\sigma}^2{\sin}^2{\theta}}{{\Omega}(i-j)^2+8(K+1){\sigma}^2{\sin}^2{\theta}} -K\right) \Bigg)}^{L_a} \text{,}
\end{split}
\end{equation}
which is monotonically increasing proved in Appendix \ref{monotony}, and that
\begin{equation} \label{eq63}
\mathscr{F}_{\text{Rician}} \left(L_a \text{,} \sigma \right) = \sum_{r=1}^{N} b_r\mathcal{F}_{\text{Rician}}(\theta_r \text{;} \sigma\text{,}L_a) \text{,}
\end{equation}
we can get the explicit bound.
\end{appendices}
\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{2011Spinal}
J.~Perry, H.~Balakrishnan, and D.~Shah, ``{Rateless Spinal Codes},''
\emph{Proceedings of the 10th ACM Workshop on Hot Topics in Networks}, 2011.
\bibitem{balakrishnan2012randomizing}
H.~Balakrishnan, P.~Iannucci, J.~Perry, and D.~Shah, ``{De-randomizing shannon:
The design and analysis of a capacity-achieving rateless code},'' \emph{arXiv
preprint arXiv:1206.0418}, 2012.
\bibitem{2012Spinal}
J.~Perry, P.~A. Iannucci, K.~E. Fleming, H.~Balakrishnan, and D.~Shah,
``{Spinal Codes},'' \emph{Proceedings of the ACM SIGCOMM 2012 Conference},
pp. 49–--60, 2012.
\bibitem{TWC1}
G.~Ungerboeck, ``{Channel coding with multilevel/phase signals},'' \emph{IEEE
Transactions on Information Theory}, vol.~28, no.~1, pp. 55--67, 1982.
\bibitem{TWC2}
G.~Ungerboeck and I.~Csajka, ``{On Improving data-link performance by
increasing the channel alphabet and introducing sequence coding},''
\emph{IEEE International Symposium on Information Theory}, 1976.
\bibitem{NBLLBC}
S.~Yousefi and A.~Khandani, ``{A new upper bound on the ML decoding error
probability of linear binary block codes in AWGN interference},'' \emph{IEEE
Transactions on Information Theory}, vol.~50, no.~12, pp. 3026--3036, 2004.
\bibitem{polar1}
D.~Goldin and D.~Burshtein, ``{Performance Bounds of Concatenated Polar Coding
Schemes},'' \emph{IEEE Transactions on Information Theory}, vol.~65, no.~11,
pp. 7131--7148, 2019.
\bibitem{polar2}
B.~Shuval and I.~Tal, ``{A Lower Bound on the Probability of Error of Polar
Codes over BMS Channels},'' \emph{IEEE Transactions on Information Theory},
vol.~65, no.~4, pp. 2021--2045, 2019.
\bibitem{raptorerror}
F.~Lázaro, G.~Liva, G.~Bauch, and E.~Paolini, ``{Bounds on the Error
Probability of Raptor Codes Under Maximum Likelihood Decoding},'' \emph{IEEE
Transactions on Information Theory}, vol.~67, no.~3, pp. 1537--1558, 2021.
\bibitem{LTerror1}
B.~Schotsch, G.~Garrammone, and P.~Vary, ``{Analysis of LT Codes over Finite
Fields under Optimal Erasure Decoding},'' \emph{IEEE Communications Letters},
vol.~17, no.~9, pp. 1826--1829, 2013.
\bibitem{LTerror2}
B.~Schotsch, \emph{{Rateless coding in the finite length regime}}.\hskip 1em
plus 0.5em minus 0.4em\relax Hochschulbibliothek der
Rheinisch-Westf{\"a}lischen Technischen Hochschule Aachen, 2014.
\bibitem{CIT-009}
I.~Sason and S.~Shamai, ``{Performance Analysis of Linear Codes under
Maximum-Likelihood Decoding: A Tutorial},'' \emph{Foundations and Trends in
Communications and Information Theory}, vol.~3, no. 1–2, pp. 1--222, 2006.
\bibitem{UEPspinal}
X.~Yu, Y.~Li, W.~Yang, and Y.~Sun, ``{Design and Analysis of Unequal Error
Protection Rateless Spinal Codes},'' \emph{IEEE Transactions on
Communications}, vol.~64, no.~11, pp. 4461--4473, 2016.
\bibitem{polyanskiy2010channel}
Y.~Polyanskiy, H.~V. Poor, and S.~Verd{\'u}, ``{Channel coding rate in the
finite blocklength regime},'' \emph{IEEE Transactions on Information Theory},
vol.~56, no.~5, pp. 2307--2359, 2010.
\bibitem{gallager1968information}
R.~G. Gallager, \emph{{Information theory and reliable communication}}.\hskip
1em plus 0.5em minus 0.4em\relax Wiley, 1968, vol. 588.
\bibitem{li2021new}
A.~Li, S.~Wu, J.~Jiao, N.~Zhang, and Q.~Zhang, ``{New Upper Bounds on the Error
Probability under ML Decoding for Spinal Codes and the Joint
Transmission-Decoding System Design},'' \emph{arXiv preprint
arXiv:2101.07953}, 2021.
\bibitem{li2021spinal}
A.~\vspace{0mm}Li, S.~Wu, J.~Jiao, N.~Zhang, and Q.~Zhang, ``{Spinal Codes Over
Fading Channel: Error Probability Analysis and Encoding Structure
Improvement},'' \emph{IEEE Transactions on Wireless Communications}, vol.~20,
no.~12, pp. 8288--8300, 2021.
\bibitem{chernoff1952measure}
H.~Chernoff, ``{A Measure of Asymptotic Efficiency for Tests of a Hypothesis
Based on the sum of Observations},'' \emph{The Annals of Mathematical
Statistics}, pp. 493--507, 1952.
\bibitem{boole1847mathematical}
G.~Boole, \emph{{The mathematical analysis of logic}}.\hskip 1em plus 0.5em
minus 0.4em\relax Philosophical Library, 1847.
\bibitem{Qfunction}
U.~Erez, M.~D. Trott, and G.~W. Wornell, ``{Rateless Coding for Gaussian
Channels},'' \emph{IEEE Transactions on Information Theory}, vol.~58, no.~2,
pp. 530--547, 2012.
\bibitem{hashfunction}
M.~Mitzenmacher and E.~Upfal, \emph{{Probability and computing: Randomized
algorithms and probabilistic analysis}}.\hskip 1em plus 0.5em minus
0.4em\relax Cambridge University Press, 2005.
\bibitem{Jacobian}
W.~Kaplan, \emph{{Advanced calculus}}.\hskip 1em plus 0.5em minus 0.4em\relax
Pearson Education India, 1952.
\end{thebibliography}