A novel low-complexity weighted symbol-flipping algorithm with flipping patterns to decode nonbinary low-density parity-check codes is proposed. The proposed decoding procedure updates the hard-decision received symbo...A novel low-complexity weighted symbol-flipping algorithm with flipping patterns to decode nonbinary low-density parity-check codes is proposed. The proposed decoding procedure updates the hard-decision received symbol vector iteratively in search of a valid codeword in the symbol vector space. Only one symbol is flipped in each iteration, and symbol flipping function, which is employed as the symbol flipping metric, combines the number of failed checks and the reliabilities of the received bits and calculated symbols. A scheme to avoid infinite loops and select one symbol to flip in high order Galois field search is also proposed. The design of flipping pattern＇s order and depth, which is dependent of the computational requirement and error performance, is also proposed and exemplified. Simulation results show that the algorithm achieves an appealing tradeoff between performance and computational requirement over relatively low Galois field for short to medium code length.展开更多
针对多元低密度奇偶校验(LDPC)码译码复杂度高、时延大等问题,提出了一种基于硬信息的低复杂度多元LDPC译码算法。来自信道的接收信号在初始化时,先进行非均匀量化预处理。在迭代过程中,校验节点端只需传输单个比特的二进制硬可靠度信...针对多元低密度奇偶校验(LDPC)码译码复杂度高、时延大等问题,提出了一种基于硬信息的低复杂度多元LDPC译码算法。来自信道的接收信号在初始化时,先进行非均匀量化预处理。在迭代过程中,校验节点端只需传输单个比特的二进制硬可靠度信息至变量节点。在变量节点端,可靠度信息按比特位进行简单的累加和更新,无需任何的系数修正操作。同时,变量节点使用了全信息的方式将信息传输至与其相邻的校验节点。仿真结果显示,与基于比特可靠度(BRB)的多元LDPC译码算法相比,提出的算法在较低量化比特情况下,能获得约0.3 d B的译码性能增益,且译码复杂度更低。展开更多
提出一种简单高效的GF(q)-LDPC码译码算法,将对数似然比和积译码算法中的雅可比对数利用一阶泰勒级数近似,从而降低译码时校验点计算的复杂度.与目前广泛应用的O ffset m in-sum算法相比较,在BER为10-4处性能有0.2 dB左右的提升,并且本...提出一种简单高效的GF(q)-LDPC码译码算法,将对数似然比和积译码算法中的雅可比对数利用一阶泰勒级数近似,从而降低译码时校验点计算的复杂度.与目前广泛应用的O ffset m in-sum算法相比较,在BER为10-4处性能有0.2 dB左右的提升,并且本算法中的参数设计独立于有限域的阶数.展开更多
文摘A novel low-complexity weighted symbol-flipping algorithm with flipping patterns to decode nonbinary low-density parity-check codes is proposed. The proposed decoding procedure updates the hard-decision received symbol vector iteratively in search of a valid codeword in the symbol vector space. Only one symbol is flipped in each iteration, and symbol flipping function, which is employed as the symbol flipping metric, combines the number of failed checks and the reliabilities of the received bits and calculated symbols. A scheme to avoid infinite loops and select one symbol to flip in high order Galois field search is also proposed. The design of flipping pattern＇s order and depth, which is dependent of the computational requirement and error performance, is also proposed and exemplified. Simulation results show that the algorithm achieves an appealing tradeoff between performance and computational requirement over relatively low Galois field for short to medium code length.
文摘针对多元低密度奇偶校验(LDPC)码译码复杂度高、时延大等问题,提出了一种基于硬信息的低复杂度多元LDPC译码算法。来自信道的接收信号在初始化时,先进行非均匀量化预处理。在迭代过程中,校验节点端只需传输单个比特的二进制硬可靠度信息至变量节点。在变量节点端,可靠度信息按比特位进行简单的累加和更新,无需任何的系数修正操作。同时,变量节点使用了全信息的方式将信息传输至与其相邻的校验节点。仿真结果显示,与基于比特可靠度(BRB)的多元LDPC译码算法相比,提出的算法在较低量化比特情况下,能获得约0.3 d B的译码性能增益,且译码复杂度更低。
