Document Type



Doctor of Philosophy


Electrical Engineering

First Adviser

Yan, Zhiyuan

Other advisers/committee members

Yan, Zhiyuan; Wagh, Meghanad D.; Li, Tiffany Jing; Wang, Zhongfeng


Due to the rapid development of the information industry, modern communication and storage systems require much higher data rates and reliability to server various demanding applications. However, these systems suffer from noises from the practical channels. Various error correction codes (ECCs), such as Reed-Solomon (RS) codes, convolutional codes, turbo codes, Low-Density Parity-Check (LDPC) codes and so on, have been adopted in lots of current standards. With the increasing data rate, the research of more advanced ECCs and the corresponding efficient decoders will never stop.Binary LDPC codes have been adopted in lots of modern communication and storage applications due their superior error performance and efficient hardware decoder implementations. Non-binary LDPC (NB-LDPC) codes are an important extension of traditional binary LDPC codes. Compared with its binary counterpart, NB-LDPC codes show better error performance under short to moderate block lengths and higher order modulations. Moreover, NB-LDPC codes have lower error floor than binary LDPC codes. In spite of the excellent error performance, it is hard for current communication and storage systems to adopt NB-LDPC codes due to complex decoding algorithms and decoder architectures. In terms of hardware implementation, current NB-LDPC decoders need much larger area and achieve much lower data throughput.Besides the recently proposed NB-LDPC codes, polar codes, discovered by Ar{\i}kan, appear as a very promising candidate for future communication and storage systems. Polar codes are considered as a major breakthrough in recent coding theory society. Polar codes are proved to be capacity achieving codes over binary input symmetric memoryless channels. Besides, polar codes can be decoded by the successive cancelation (SC) algorithm with of complexity of $\mathcal{O}(N\log_2 N)$, where $N$ is the block length. The main sticking point of polar codes to date is that their error performance under short to moderate block lengths is inferior compared with LDPC codes or turbo codes. The list decoding technique can be used to improve the error performance of SC algorithms at the cost higher computational and memory complexities. Besides, the hardware implementation of current SC based decoders suffer from long decoding latency which is unsuitable for modern high speed communications.ECCs also find their applications in improving the reliability of network coding. Random linear network coding is an efficient technique for disseminating information in networks, but it is highly susceptible to errors. K\"{o}tter-Kschischang (KK) codes and Mahdavifar-Vardy (MV) codes are two important families of subspace codes that provide error control in noncoherent random linear network coding. List decoding has been used to decode MV codes beyond half distance. Existing hardware implementations of the rank metric decoder for KK codes suffer from limited throughput, long latency and high area complexity. The interpolation-based list decoding algorithm for MV codes still has high computational complexity, and its feasibility for hardware implementations has not been investigated.In this exam, we present efficient decoding algorithms and hardware decoder architectures for NB-LDPC codes, polar codes, KK and MV codes. For NB-LDPC codes, an efficient shuffled decoder architecture is presented to reduce the number of average iterations and improve the throughput. Besides, a fully parallel decoder architecture for NB-LDPC codes with short or moderate block lengths is also presented. Our fully parallel decoder architecture achieves much higher throughput and area efficiency compared with the state-of-art NB-LDPC decoders. For polar codes, a memory efficient list decoder architecture is first presented. Based on our reduced latency list decoding algorithm for polar codes, a high throughput list decoder architecture is also presented. At last, we present efficient decoder architectures for both KK and MV codes.