中文摘要：

低密度校验（LDPC）码最小和译码算法在BSC信道下具有较高的错误平层。针对这一问题，提出了一种改进的最小和译码算法。该算法在每轮迭代译码过程中利用变量节点邻居传入的可靠性强度阶和原始信道接收值之间存在特定的规律进行消息更新，从而能够加速译码收敛速度，提高译码性能。仿真结果表明，该算法在BSC信道下使用3或4级强度阶可以逼近甚至超越浮点型BP译码算法。

英文摘要：

Compared with the BP decoding algorithm, the min-sum decoding algorithm has higher error floor for LDPC codes over the binary symmetric channel （BSC）. In order to solve this problem, a modified algorithm is presented. The proposed algorithm can reduce memory overhead and accelerate the convergence of decoding algo- rithm by using the thought of multi-level quantification and considering the knowledge both extra messages for each node and received values from the channel, respectively. Simulation results show that the improved min-sum algo- rithm with three or four levels approach and even surpass floating-point BP decoding algorithm in frame error rate （FER） performance over the BSC.