High-Speed Decoders for Polar Codes

Abstract

Error detection and correction plays a vital role in modern information storage and communication systems. Polar codes are gathering a lot of attention as they are a class of capacity-achieving error-correcting codes with an explicit construction that can be decoded with low-complexity algorithms. However, their adoption is hindered by the lack of high-speed – high throughput and low latency – hardware and software decoders for codes of practical length and rate. This thesis presents various solutions to this problem. It introduces modifications to the state-of-the-art low-complexity decoding algorithm to better accommodate low-rate polar codes. It also proposes a code construction alteration process. Hardware implementation results show good latency reduction and throughput improvement with little to negligible coding loss for low-rate moderate-length polar codes. Then, it presents high-speed software polar decoders. It shows how adapting the decoding algorithm at various levels can lead to significant improvements in latency and throughput, yielding polar decoders that are suitable for high-performance software-defined radio applications on modern desktop processors and embed\-ded-plat\-form processors. These proposed decoders have an order of magnitude lower latency and memory footprint compared to state-of-the-art decoders, while maintaining comparable throughput. In addition, strategies and results for implementing polar decoders on graphical processing units are presented.Next, it demonstrates that polar decoders can achieve extremely high throughput values and retain moderate complexity. It presents a family of architectures for hardware polar decoders that employ unrolling. The resulting fully-unrolled architectures are capable of achieving a throughput that is two to three orders of magnitude greater than current state of the art while maintaining good energy efficiency. Moreover, the proposed architectures are flexible in a way that makes it possible to explore the trade-off between area, throughput and energy efficiency. Lastly, while unrolled decoders provide the greatest decoding speed, they are built for a specific, fixed, code i.e. the code length or rate cannot be modified at execution time. Most modern wireless communication applications largely benefit from the support of multiple code lengths and rates. This thesis shows how an unrolled decoder can be transformed into a multi-mode decoder supporting many codes of various lengths and rates. Implementation results show a peak information throughput that is an order of magnitude greater than the state of the art, while showing the best area and energy efficiency.%%%%La detection et la correction des erreurs jouent un role essentiel dans les systemes modernes de stockage et de communication. Les codes polaires intriguent actuellement beaucoup de chercheurs car ils constituent une classe de codes correcteurs capables d'atteindre la capacite theorique d'un canal avec des algorithmes de decodage de faible complexite tout en proposant une methode de…