Abstract:
A class of product-matrix regenerating codes for fault-tolerance protection in a distributed storage system is provided, designed for enhanced network bandwidth and computational performance. The codes are based on the Vandermonde matrix and their equivalent systematic codes are sparse, allowing throughputs compatible with current practical deployments of erasure codes while presenting significant improvement in network bandwidth (network related repair cost). Methods and apparatuses for encoding, regenerating and decoding data are also provided utilizing the codes.
Abstract:
The present invention relates to a method and a device for an error correction of transmitted data. For this purpose, the transmitted data are encoded in a block code, wherein the block code comprises a number of data bits and an additional number of redundant bits. Herein the block code is described by a parity-check matrix H, wherein columns of the parity-check matrix Hare inherently related to the data bits of the block code. The method according to the present invention comprises the following steps: (a) diagonalizing the parity-check matrix H, with respect to at least one column of the parity-check matrix H,into a diagonalized parity-check matrix H', wherein the diagonalized parity-check matrix H' is related to the block code and to the at least one column; (b) determining at least one error position (130) in the block code by using the diagonalized parity-check matrix H' and a syndrome vector, wherein the syndrome vector is related to the data bits in the block code; (c) performing the error correction of the transmitted data at the at least one error position (130) in the block code. The present method and device allow providing communication channels with increased reliability and enhanced correction capability at reduced complexity,and is generally applicable to all known block codes, such as turbo, LDPC, BCH, or Reed-Solomon codes.
Abstract:
本发明实施例提供一种极性码的译码方法及装置。所述方法包括:获取由输入映射序列进行编码后在信道上输出的接收序列和校验矩阵;根据接收序列和校验矩阵进行Trellis路径搜索,并计算输入映射序列对应的Trellis路径似然信息;根据输入映射序列对应的Trellis路径似然信息计算输入映射序列对应的判决值;按预设的映射关系对输入映射序列对应的判决值进行解映射,得到译码序列。输入映射序列为按预设的映射关系对输入序列进行映射得到的序列,输入映射序列从属于有限域GF(q),q=2 m ,m为大于1的正整数,在进行极性码译码时,进行Trellis路径搜索,计算判决值,再解映射,可同时降低最大list值,且大幅降低平均list值。
Abstract:
A method for generating a codebook includes applying a unitary rotation to a baseline multidimensional constellation to produce a multidimensional mother constellation, wherein the unitary rotation is selected to optimize a distance function of the multidimensional mother constellation, and applying a set of operations to the multidimensional mother constellation to produce a set of constellation points. The method also includes storing the set of constellation points as the codebook of the plurality of codebooks.
Abstract:
A method of data encoding is disclosed. A communications device receives a set of information bits to be encoded into a codeword (c), which includes the set of information bits and a set of parity bits. A first bit value is assigned to a first parity bit in the set of parity bits. The remaining parity bits are then encoded based, at least in part, on the first bit value assigned to the first parity bit. The device may encode the remaining parity bits using the set of information bits and a parity check matrix (H) for a low density parity check (LDPC) code. The device may also generate a new parity check matrix (H 0 ) based on linearly independent rows of the parity check matrix H, and iteratively evaluate each of the remaining parity bits based on the equation: H 0 c = 0. The device may then determine whether the encoded codeword c is a valid codeword given the LDPC code, and change one or more bit values of the codeword if c is not a valid codeword.