Channel Coding for Gaussian Channels with Mean and Variance Constraints

Adeel Mahmood Aaron B. Wagner Nokia Bell Labs School of Electrical and Computer Engineering, Cornell University · 2025

Read Channel Coding for Gaussian Channels with Mean and Variance Constraints by Adeel Mahmood Aaron B. Wagner Nokia Bell Labs School of Electrical and Computer Engineering, Cornell University (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-10953.txt.

Introduction

Adeel Mahmood and Aaron Wagner pin down the best codes for a noisy channel with two caps on the signal in a 2025 paper. The caps bind the mean and the spread of what is sent, a fresh twist on a classic setup. The team match a no-go bound with a build that meets it, for both the main rate and the next-order term.

The build draws code words from a blend of three ball shells with near-equal radii. A key lemma caps how far the output laws of two shells can drift, with high odds. That cap lets the classic bell-curve tools do the rest.

Catalog record

CREATORAdeel Mahmood Aaron B. Wagner Nokia Bell Labs School of Electrical and Computer Engineering, Cornell University
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE94 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-10953

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Channel Coding for Gaussian Channels with Mean and Variance Constraints Adeel Mahmood Aaron B. Wagner Nokia Bell Labs School of Electrical and Computer Engineering, Cornell University Abstract We consider channel coding for Gaussian channels with the recently introduced mean and variance cost constraints. Through matching converse and achievability bounds, we characterize the optimal first- and second-order performance. The main technical contribution of this paper is an achievability scheme which uses random codewords drawn from a mixture of three uniform distributions on ( n − 1 ) (n-1) -spheres of radii R 1 , R 2 R_{1},R_{2} and R 3 R_{3} , where R i = O ⁡ ( n ) R_{i}=O(\sqrt{n}) and | R i − R j | = O ⁡ ( 1 ) |R_{i}-R_{j}|=O(1) . To analyze such a mixture distribution, we prove a lemma giving a uniform O ⁡ ( log ⁡ n ) O(\log n) bound, which holds with high probability, on the log ratio of the output distributions Q i c ​ c Q_{i}^{cc} and Q j c ​ c Q_{j}^{cc} , where Q i c ​ c Q_{i}^{cc} is induced by a random channel input uniformly distributed on an ( n − 1 ) (n-1) -sphere of radius R i R_{i} . To facilitate the application of the usual central limit theorem, we also give a uniform O ⁡ ( log ⁡ n ) O(\log n) bound, which holds with high probability, on the log ratio of the output distributions Q i c ​ c Q_{i}^{cc} and Q i ∗ Q^{*}_{i} , where Q i ∗ Q_{i}^{*} is induced by a random channel input with i.i.d. components. I Introduction The two common forms of cost (or power) constraints in channel coding have been the maximal cost constraint specified by c(\mathbf{X})\leq\Gamma almost surely, or the expected cost constraint specified by \mathbb{E}\left[c(\mathbf{X})\right]\leq\Gamma , where \mathbf{X}\in\mathbb{R}^{n} is a random channel input vector and c(\cdot) is an additively separable cost function defined as \displaystyle c(\mathbf{X})\coloneqq\frac{1}{n}\sum_{i=1}^{n}c(X_{i}). Recent works [ 1 ] and [ 2 ] introduced the mean and variance (m.v.) cost constraint, specified by \displaystyle\begin{split}\mathbb{E}\left[c(\mathbf{X})\right]&\leq\Gamma\\ \text{Var}\left(c(\mathbf{X})\right)&\leq\frac{V}{n},\end{split} (1) for discrete memoryless channels (DMCs). With a variance parameter V , the mean and variance (m.v.) cost constraint generalizes the two existing frameworks in the sense that V\to 0 recovers the first- and second-order coding performance of the maximal cost constraint [ 1 ] and V\to\infty recovers the expected cost constraint. Beyond generalization, the m.v. cost constraint for 0<V<\infty is shown to have practical advantages over both prior cost models. Unlike the maximal cost constraint, it allows for an improved second-order coding performance with feedback ( [ 1 , Theorem 3] , [ 2 , Theorem 2] ). In particular, for DMCs with a unique capacity-cost-achieving distribution, second-order coding performance improvement via feedback is possible if and only if V>0 . Even without feedback, the coding performance under the m.v. cost constraint is superior ( [ 2 , Theorem 1] ). Specifically, if r(V) and r(0) denote the optimal non-feedback second-order coding rates under the m.v. and maximal cost constraints, respectively, then [2, Theorem 1] showed that r(V)>r(0) for all V>0 . Unlike the expected cost constraint, the m.v. cost constraint enforces a controlled, “ergodic” use of transmission power [ 2 , (5)] . This is essential for several practical reasons such as operating circuitry in the linear regime, minimizing power consumption, and reducing interference with other terminals. The m.v. cost constraint allows the power to fluctuate above the threshold in a manner consistent with a noise process, thus making it a more realistic and natural cost model in practice than the restrictive maximal cost constraint. The aforementioned performance improvements under the m.v. cost constraint were shown for DMCs only, although the ergodic behavior enforced by the m.v. cost constraint is generally true. In this paper, we study additive white Gaussian noise (AWGN) channels with the mean and variance cost constraint with the cost function taken to be c(x)=x^{2} . Specifically, we characterize the optimal first- and second-order coding rates (SOCR) under the m.v. cost constraint for AWGN channels subject to a non-vanishing average probability of error \epsilon>0 . Simultaneously, we also characterize the optimal average error probability as a function of the second-order rate with the baseline first-order rate fixed as the capacity-cost function [ 3 , (9.17)] . The latter is motivated by the fact, which itself follows from our results, that the strong converse [ 3 , p. 208] holds under the m.v. cost constraint. Concretely, we show that the asymptotic expansion of the maximum achievable rate R^{*}(n,\epsilon,\Gamma,V) under the m.v. cost constraint is given by \displaystyle R^{*}(n,\epsilon,\Gamma,V)=C(\Gamma)+O\left(\frac{1}{\sqrt{n}}\right), (2) where C(\Gamma) denotes the capacity-cost function. The fact that the first-order term C(\Gamma) is independent of \epsilon\in(0,1) is an interesting finding since the strong converse does not hold for AWGN channels subject to expectation-only constraint [ 4 , Theorem 77] . However, if one considers maximal probability of error instead of average probability of error, then the strong converse holds for both m.v. cost constraint and expectation-only constraint [ 5 , (18)] . In this paper, we only focus on the average error probability and non-feedback framework. Another interesting observation about (\ref{3m}) is that the first-order term C(\Gamma) is independent of the variance parameter V for all finite V\geq 0 . Instead, the variance parameter along with the channel dispersion plays a role in the second-order term O(1/\sqrt{n}) , whose exact characterization is discussed later in the paper. More results on AWGN channels can be found in [ 6 ] and [ 7 ] . It is also to be noted that the asymptotic expansion in (\ref{3m}) is proven for a fixed finite variance parameter V\geq 0 as n\to\infty . Therefore, while V\to\infty in (\ref{43v}) recovers the expectation-only constraint, we have \displaystyle C(\Gamma)=\lim_{V\to\infty}\lim_{n\to\infty}R^{*}(n,\epsilon,\Gamma,V)\neq\lim_{n\to\infty}\lim_{V\to\infty}R^{*}(n,\epsilon,\Gamma,V)\stackrel{{\scriptstyle(a)}}{{=}}C\left(\frac{\Gamma}{1-\epsilon}\right), (3) where the equality (a) above can be found in [ 4 , Theorem 77] . The achievability scheme in [ 4 , Theorem 77] achieves \lim_{n\to\infty}\lim_{V\to\infty}R^{*}(n,\epsilon,\Gamma,V)\geq C\left(\frac{\Gamma}{1-\epsilon}\right) by assigning zero cost to an approximately \epsilon fraction of the codewords while drawing the remaining codewords uniformly from a sphere of radius approximately equal to \sqrt{\frac{n\Gamma}{1-\epsilon}} . Such a scheme has \operatorname{Var}(c(\mathbf{X}))=O(1) and is disallowed under the m.v. cost constraint formulation in (\ref{43v}) . Moreover, in contrast to the O(1/\sqrt{n}) second-order term in (\ref{3m}) , we have \displaystyle R^{*}(n,\epsilon,\Gamma,\infty)=C\left(\frac{\Gamma}{1-\epsilon}\right)+O\left(\sqrt{\frac{\log n}{n}}\right) (4) from [ 5 , Theorem 1] . Hence, the analysis under the m.v. constraint with a finite variance parameter V\geq 0 requires different techniques. Our achievability proof relies on a random coding scheme where the channel input has a distribution supported on at most three concentric spheres of radii R_{1} , R_{2} and R_{3} , where R_{i}=O(\sqrt{n}) and |R_{i}-R_{j}|=O(1) . Each uniform distribution on a sphere induces a distribution Q_{i}^{cc} on the channel output, and a mixture of such input uniform distributions induces a mixture distribution of Q_{1}^{cc},Q_{2}^{cc} and Q_{3}^{cc} on the output. For \mathbf{Y}\sim Q_{i}^{cc} , we have that \mathbb{E}[||\mathbf{Y}||^{2}]=O(n) and ||\mathbf{Y}||^{2} concentrates in probability over a set encompassing \pm\sqrt{n} deviations around its mean. Furthermore, the probability density function Q_{i}^{cc} is given in terms of the modified Bessel function of the first kind. To assist in the analysis of this mixture distribution, we use a

Read the complete text