Channel Coding Rate in the Finite Blocklength Regime

被引：2468

作者：

Polyanskiy, Yury ^{[1
]}

Poor, H. Vincent ^{[1
]}

Verdu, Sergio ^{[1
]}

机构：

[1] Princeton Univ, Dept Elect Engn, Princeton, NJ 08544 USA

来源：

IEEE TRANSACTIONS ON INFORMATION THEORY | 2010年 / 56卷 / 05期

基金：

美国国家科学基金会;

关键词：

Achievability; channel capacity; coding for noisy channels; converse; finite blocklength regime; Shannon theory; ERROR-PROBABILITY; RANDOM CODES; BOUNDS;

D O I：

10.1109/TIT.2010.2043769

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

This paper investigates the maximal channel coding rate achievable at a given blocklength and error probability. For general classes of channels new achievability and converse bounds are given, which are tighter than existing bounds for wide ranges of parameters of interest, and lead to tight approximations of the maximal achievable rate for blocklengths n as short as 100. It is also shown analytically that the maximal rate achievable with error probability c is closely approximated by C -root V/n Q(-1()c) where C is the capacity, V is a characteristic of the channel referred to as channel dispersion, and Q is the complementary Gaussian cumulative distribution function.

引用

页码：2307 / 2359

页数：53

共 51 条

[1] Abramowitz M., 1972, Handbook on Mathematical Functions with Formulas, Graphs, and Mathematical Tables
[2] Ash Robert B., 1965, Information theory
[3] ASHIKHMIN A, 2009, COMMUNICATION
[4] A new upper bound on the reliability function of the Gaussian channel
Ashikhmin, AE
Barg, A
Litsyn, SN
[J]. IEEE TRANSACTIONS ON INFORMATION THEORY, 2000, 46 (06) : 1945 - 1961
[5] Random codes: Minimum distances and error exponents
Barg, A
Forney, GD
[J]. IEEE TRANSACTIONS ON INFORMATION THEORY, 2002, 48 (09) : 2568 - 2573
[6] BARON D, 2004, P 38 AS C SIGN SYST
[7] HYPOTHESIS TESTING AND INFORMATION-THEORY
BLAHUT, RE
[J]. IEEE TRANSACTIONS ON INFORMATION THEORY, 1974, 20 (04) : 405 - 417
[8] BUCKINGHAM D, 2008, P C INF SCI SYST CIS
[9] Csiszar I, 1967, Stud Sci Math Hung, V2, P229
[10] Di CY, 2002, IEEE T INFORM THEORY, V48, P1570, DOI 10.1109/TIT.2002.1003839

← 1 2 3 4 5 6 →