Performance of Polar Codes for Channel and Source Coding

TL;DR

This paper reviews the performance of polar codes in channel and source coding, highlighting their capacity-achieving properties, finite-length improvements, and applications in lossless, lossy, and distributed scenarios.

Contribution

It discusses techniques to enhance finite-length performance of polar codes and explores their effectiveness in various source coding contexts.

Findings

Polar codes achieve capacity with low complexity decoding.

Finite-length performance can be improved with specific techniques.

Polar codes are effective in both channel and source coding applications.

Abstract

Polar codes, introduced recently by Ar\i kan, are the first family of codes known to achieve capacity of symmetric channels using a low complexity successive cancellation decoder. Although these codes, combined with successive cancellation, are optimal in this respect, their finite-length performance is not record breaking. We discuss several techniques through which their finite-length performance can be improved. We also study the performance of these codes in the context of source coding, both lossless and lossy, in the single-user context as well as for distributed applications.