Partial geometries form an interesting branch in combinatorial mathematics, and recently they have been shown to be very effective in the construction of LDPC codes with distinct geometric and algebraic structures. This paper presents three specific cyclic classes of partial geometries. Based on these three classes of partial geometries, three classes of LDPC codes and three classes of constant-weight codes are constructed. Codes in these two categories are either cyclic or quasi-cyclic. Designs and constructions of these codes are straightforward and flexible without a need for extensive computer search. It is shown that long high-rate LDPC codes constructed based on the three classes of cyclic partial geometries perform well over the additive white Gaussian noise channel (AWGNC) with iterative decoding algorithm based on belief propagation. They can achieve low error-rates without visible error-floor and their decoding converges rapidly. These LDPC codes also perform well over the binary erasure channel (BEC) and are very effective in correcting phased-bursts of erasures. Based on cyclic partial geometries, a special type of constant-weight codes, called balanced codes, can be constructed. The constant-weight codes constructed based on partial geometries are either optimal or nearly optimal.

In his pioneering work on LDPC codes, Gallager dismissed codes with parity-check matrices of weight two after proving that their minimum Hamming distances grow at most logarithmically with their code lengths. In spite of their poor minimum Hamming distances, it is shown that quasi-cyclic LDPC codes with parity-check matrices of column weight two have good capability to correct phased bursts of erasures which may not be surpassed by using quasi-cyclic LDPC codes with parity-check matrices of column weight three or more. By modifying the parity-check matrices of column weight two and globally coupling them, the erasure correcting capability can be further enhanced. Quasi-cyclic LDPC codes with parity-check matrices of column weight three or more that can correct phased bursts of erasures and perform well over the AWGN channel are also considered. Examples of such codes based on Reed-Solomon and Gabidulin codes are presented.

A novel scheme is presented for encoding and iterative soft-decision decoding of cyclic codes of prime lengths. The encoding of a cyclic code of a prime length is performed on a collection of codewords which are mapped through Galois Fourier transform into a codeword in a low-density parity-check code with a binary parity-check matrix for transmission. Using this matrix, binary iterative soft-decision decoding algorithm is applied to jointly decode a collection of codewords from the cyclic code. The joint-decoding allows for information sharing among the received vectors corresponding to the codewords in the collection during the iterative decoding process. For decoding Reed-Solomon and BCH codes of prime lengths, the proposed decoding scheme not only requires much lower decoding complexity than other soft-decision decoding algorithms for these codes, but also yields superior performance. The proposed decoding scheme can also achieve a joint-decoding gain over the maximum likelihood decoding of individual codewords. The decoding scheme is also applied to quadratic residue codes.

This paper presents a construction of a class of partial geometries based on RS codes of prime lengths and shows that LDPC codes constructed based on Reed-Solomon codes of prime lengths are finite geometry LDPC codes. Furthermore, a new method for design and construction of nonbinary quasi-cyclic LDPC codes based on the conventional parity-check matrices of Reed-Solomon codes is presented. Simulation results show that the constructed nonbinary LDPC codes perform well over the additive white Gaussian channel.

This paper presents a special type of nonbinary LDPC codes which are constructed based on Reed-Solomon codes. For a code of this type, its Tanner graph is composed of a set of disjoint and identical Tanner graphs, which are coupled together by a group of global check-nodes. Such a code is called a globally coupled LDPC code. This type of codes are capable of correcting random symbol errors, multiple phased bursts of erasures, and a single long burst of erasures.

A novel scheme is presented for encoding and decoding of Reed-Solomon codes of prime lengths. Encoding is performed on a collection of codewords which are mapped through Galois Fourier transform into a codeword in a low-density parity-check code with a binary parity-check matrix for transmission. Using this matrix, a binary iterative soft-decision decoding algorithm is applied to jointly decode a collection of codewords in the Reed-Solomon code. By allowing information sharing among the received vectors corresponding to the code-words in the collection, the proposed decoding scheme achieves superior performance over algorithms decoding individual Reed-Solomon codewords including maximum likelihood decoding.

The use of partial geometries to construct parity-check matrices for binary low-density parity-check (LDPC) codes has resulted in the design of successful codes with a probability of error on the AWGN channel close to the Shannon capacity at bit error rate down to $10^{-15}$ . Such considerations have motivated this further investigation. A new and simple construction of a type of partial geometries with a quasi-cyclic (QC) structure is given and their properties are investigated. Two new classes of this type of partial geometries, one based on prime fields and the other based on cyclic subgroups of prime orders of finite fields, are constructed. QC-LDPC codes with good error performances are constructed based on these two new classes of partial geometries. The trapping sets of the partial geometry codes were previously considered using the geometric aspects of the underlying structure to derive information on the size of allowable trapping sets. This topic is further considered here. Finally, there is a natural relationship between partial geometries and strongly regular graphs. The eigenvalues of the adjacency matrices of such graphs are well known, and it is of interest to determine if any of the Tanner graphs derived from the partial geometries are good expanders for certain parameter sets, since it can be argued that codes with good geometric and expansion properties might perform well on the AWGN channel under message-passing decoding.

Recently, Karimi and Banihashemi proposed an algorithm based on message-passing to count cycles in a graph of lengths less than double its girth. The algorithm uses only integer additions and subtractions to compute messages at the nodes of the graph that are passed to adjacent nodes. The complexity of the algorithm, when applied to a bipartite graph of girth g that has E edges, is O(gE2). The algorithm is superior to many other existing algorithms in the literature. In this paper, an improvement of this algorithm is presented that cuts both the complexity and the computing time by a factor of two. The improved algorithm is also applied to Tanner graphs of quasi-cyclic codes and, in this case, the complexity can be further cut by a factor of p, where p is the size of the circulants in the parity-check matrix of the quasi-cyclic code.

This paper presents two simple and very flexible methods for constructing non-binary (NB) quasi-cyclic (QC) LDPC codes. The proposed construction methods have several known ingredients including base array, masking, binary to nonbinary replacement, and matrix-dispersion. By proper choice and combination of these ingredients, NB-QC-LDPC codes with excellent performance can be constructed. The constructed codes can be decoded with a reduced-complexity iterative decoding scheme which significantly reduces the hardware implementation complexity.

This paper presents a reduced-complexity iterative scheme and an algorithm for decoding non-binary quasi-cyclic (QC) LDPC codes of a specific type. The proposed decoding scheme and the algorithm together significantly reduce the hardware implementation complexity of a decoder with no performance degradation. Also presented in the paper is a simple method for constructing a class of non-binary QC-LDPC codes.

This paper presents a simple and flexible method for constructing QC-LDPC codes based on two arbitrary sets of a finite field. Based on this method, a high-rate, high-performance and very low error-floor QC-LDPC code is first constructed and then a class of rate-1/2 QC-LDPC codes whose Tanner graphs have girth 8 or larger is presented. Also presented is a reduced-complexity iterative decoding algorithm for QC-LDPC codes.

This paper presents a simple and very flexible method for constructing quasi-cyclic (QC) low density paritycheck (LDPC) codes based on finite fields. The code construction is based on two arbitrary subsets of elements from a given field. Some well known constructions of QC-LDPC codes based on finite fields and combinatorial designs are special cases of the proposed construction. The proposed construction in conjunction with a technique, known as masking, results in codes whose Tanner graphs have girth 8 or larger. Experimental results show that codes constructed using the proposed construction perform well and have low error-floors. Also presented in the paper is a reduced-complexity iterative decoding scheme for QC-LDPC codes based on the section-wise cyclic structure of their parity-check matrices. The proposed decoding scheme is an improvement of an earlier proposed reduced-complexity iterative decoding scheme.