Yu Kou

Low-density parity-check codes based on finite geometries: a rediscovery and new results

This paper presents a geometric approach to the construction of low-density parity-check (LDPC) codes. Four classes of LDPC codes are constructed based on the lines and points of Euclidean and projective geometries over finite fields. Codes of these four classes have good minimum distances and their Tanner (1981) graphs have girth 6. Finite-geometry LDPC codes can be decoded in various ways, ranging from low to high decoding complexity and from reasonably good to very good performance. They perform very well with iterative decoding. Furthermore, they can be put in either cyclic or quasi-cyclic form. Consequently, their encoding can be achieved in linear time and implemented with simple feedback shift registers. This advantage is not shared by other LDPC codes in general and is important in practice. Finite-geometry LDPC codes can be extended and shortened in various ways to obtain other good LDPC codes. Several techniques of extension and shortening are presented. Long extended finite-geometry LDPC codes have been constructed and they achieve a performance only a few tenths of a decibel away from the Shannon theoretical limit with iterative decoding.

Iterative decoding of one-step majority logic deductible codes based on belief propagation

Previously, the belief propagation (BP) algorithm has received a lot of attention in the coding community, mostly due to its near-optimum decoding for low-density parity check (LDPC) codes and its connection to turbo decoding. In this paper, we investigate the performance achieved by the BP algorithm for decoding one-step majority logic decodable (OSMLD) codes. The BP algorithm is expressed in terms of likelihood ratios rather than probabilities, as conventionally presented. The proposed algorithm fits better the decoding of OSMLD codes with respect to its numerical stability due to the fact that the weights of their check sums are often much higher than that of the corresponding LDPC codes. Although it has been believed that OSMLD codes are far inferior to LDPC codes, we show that for medium code lengths (say between 200-1000 bits), the BP decoding of OSMLD codes can significantly outperform BP decoding of their equivalent LDPC codes. The reasons for this behavior are elaborated.

Construction of low-density parity-check codes based on balanced incomplete block designs

This correspondence presents a method for constructing structured regular low-density parity-check (LDPC) codes based on a special type of combinatoric designs, known as balance incomplete block designs. Codes constructed by this method have girths at least 6 and they perform well with iterative decoding. Furthermore, several classes of these codes are quasi-cyclic and hence their encoding can be implemented with simple feedback shift registers.

On algebraic construction of Gallager and circulant low-density parity-check codes

This correspondence presents three algebraic methods for constructing low-density parity-check (LDPC) codes. These methods are based on the structural properties of finite geometries. The first method gives a class of Gallager codes and a class of complementary Gallager codes. The second method results in two classes of circulant-LDPC codes, one in cyclic form and the other in quasi-cyclic form. The third method is a two-step hybrid method. Codes in these classes have a wide range of rates and minimum distances, and they perform well with iterative decoding.

Low density parity check codes: construction based on finite geometries

Low density parity check (LDPC) codes with iterative decoding based on belief propagation (IDBP) achieve astonishing error performance close to the Shannon limit. Until now there has been no known method for constructing these Shannon limit approaching codes systematically. Good LDPC codes are largely generated by computer search. As a result, the encoding of long LDPC codes is in general very complex. This paper presents the first algebraic method for constructing LDPC codes systematically based on finite analytic geometries. Four classes of finite geometry LDPC codes with relatively good minimum distances are constructed. These codes are either cyclic or quasi-cyclic and therefore their encoding can be implemented with simple linear feedback shift registers. Long finite geometry LDPC codes have been constructed and they achieve an error performance only a few tenths of a dB away from the Shannon limit. Finite geometry LDPC codes are strong competitors to turbo codes for error control in communication and digital data storage systems.

On circulant low density parity check codes

This paper presents three classes of low density parity check codes constructed based on the cyclic structure of lines in finite geometries. Codes in these classes are either cyclic or quasi-cyclic, and they perform well with iterative decoding.

Low density parity check codes based on finite geometries: a rediscovery

LDPC codes with iterative decoding based on belief propagation have been shown to achieve astonishing error performance. But no algebraic or geometric method has been found for constructing these codes. Codes that have been found are largely computer generated, especially long codes. In this paper, we present two classes of high rate LDPC codes whose constructions are based on the lines of two-dimensional finite Euclidean and projective geometries, respectively.

Construction of low density parity check codes: a combinatoric design approach

This paper presents a method for constructing low density parity check (LDPC) codes based on a special type of combinatoric designs, known as the balanced incomplete block designs (BIBDs). Several classes of BIBDs suitable for constructing LDPC codes are presented. Codes constructed based on these classes of BIBDs perform well with iterative decoding.

A general class of LDPC finite geometry codes and their performance

We present a general class of finite geometry LDPC codes which perform well with iterative decoding although their Tanner graphs may contain many cycles of length 4. A hybrid two-stage decoding algorithm is proposed that combines iterative and multi-step majority-logic decodings to achieve good performance with low decoding complexity.

On a class of finite geometry low density parity check codes

A new class of geometry LDPC codes is presented which contains the class of Kou-Lin-Fossorier codes (see IEEE International Symposium on Information Theory, p.200, June 2000) as a subclass. If the code construction is based on Euclidean geometry (EG) and projective geometry (PG) over finite fields, we obtain four classes of LDPC codes, namely: (1) type-I EG-LDPC codes; (2) type-II EG-LDPC codes; (3) type-I PG-LDPC codes; and (4) type-II PG-LDPC codes.