CRC Aided Dynamic List SCL Decoder for Polar and PAC Codes

Abstract

Error-correction codes are indispensable to ensure reliable data transmission in various communication systems. Among these, Polar codes and PAC (Polarized Adjustable Convolutional) codes have gained significant attention due to their capacity-achieving properties and low encoding complexity. This thesis presents a novel decoding technique that combines the Targeted Parity Protection (TPP) algorithm and Dynamic Successive Cancellation List (SCL) decoding to achieve substantial performance improvements while maintaining much lower complexity compared to the conventional Fano decoder. The proposed technique focuses on leveraging Cyclic Redundancy Check (CRC) bits to protect critical information bits, leading to enhanced error correction capabilities. By applying TPP, the decoding process is accelerated by effectively reducing latency and saves computational resources by eliminating wrong codeword candidates at an earlier stage before the decoder favors them. The integration of Dynamic SCL introduces adaptability to the decoder, allowing it to adjust the list size based on the decoding requirements for different code rates and lengths. This adaptability enables efficient use of hardware resources by optimizing the utilization of the decoder storage but also tailors the decoder's performance to each specific scenario, yielding substantial error correction performance gains in both Polar and PAC codes. Extensive simulations and performance evaluations demonstrate the effectiveness of the proposed technique. The results indicate a gain of up to 0.4 dB for Polar codes and up to 0.6 dB for PAC codes compared to the standard SCL decoding approach. It is worth mentioning that such gains are considered substantial for fiber optical communications which is one of the target applications of the new technique. Furthermore, the technique showcases a reduction in decoding complexity, outperforming the traditional Fano decoder.