High-Throughput Polar Code SCL Decoding with Dynamically Configurable Pipeline Architecture

Abstract

Due to the low complexity and outstanding performance, polar code is recognized as one of the most popular channel coding in wireless communication. However, the conventional successive cancellation (SC) decoder suffers from modest performance compared with other decoders. To deal with this issue, another novel decoder, namely successive cancellation list (SCL), is developed with significant improvement in bit error ratio. However, the performance improvement is at the cost of high complexity, which leads to high decoding latency. In this paper, we propose a dynamically configurable decoding architecture with inter-frame pipelining and integrate the SCL decoding algorithm, SC decoding algorithm, and fast successive cancellation (FSC) decoding algorithm, which can reduce the complexity significantly. Compared with existing architectures, our work achieves a higher throughput and supports various codes lengths from 128 to 1024 and various code rates to meet the requirements of different applications.