Increasing the Speed of Polar List Decoders

TL;DR

This paper introduces a faster polar list decoding algorithm using a Chase-like process, achieving up to six times speed-up while maintaining error-correction performance, and proposes an adaptive method for even higher throughput.

Contribution

A simplified decoding algorithm for polar codes that significantly increases decoding speed without sacrificing error correction, and an adaptive approach for further throughput enhancement.

Findings

Six times faster than standard polar list decoding

Up to 16 times faster than LDPC decoders with similar performance

Maintains error-correction performance while increasing speed

Abstract

In this work, we present a simplified successive cancellation list decoder that uses a Chase-like decoding process to achieve a six time improvement in speed compared to successive cancellation list decoding while maintaining the same error-correction performance advantage over standard successive-cancellation polar decoders. We discuss the algorithm and detail the data structures and methods used to obtain this speed-up. We also propose an adaptive decoding algorithm that significantly improves the throughput while retaining the error-correction performance. Simulation results over the additive white Gaussian noise channel are provided and show that the proposed system is up to 16 times faster than an LDPC decoder of the same frame size, code rate, and similar error-correction performance, making it more suitable for use as a software decoding solution.