卷积码盲识别树图遍历算法研究与仿真实验

Research and simulation experiment of blind recognition tree graph traversal algorithm for convolutional codes

针对现有卷积码盲识别算法在校验矩阵相同的情况下识别性能不佳的问题,提出一种基于树图遍历的卷积码盲识别算法,并对算法识别准确率进行仿真实验。算法基本思想是:遍历所有可能的输入序列,并依据输入序列构造下一状态,根据输入序列和状态这两个参数在卷积码树图中寻找符合要求的路径,识别出卷积码的输入⁃状态⁃输出转移表,然后构造方程组,估计出生成矩阵。实验结果表明,该算法能够有效地对卷积码进行盲识别,规避了从校验矩阵反推生成矩阵的过程中存在多个可能的生成矩阵的问题。

In allusion to the poor recognition performance of the existing blind convolutional code recognition algorithm in the case of the same check matrix,a convolutional code blind recognition algorithm based on tree graph traversal is proposed in this paper,and the recognition accuracy of the algorithm is simulated.The basic idea of the algorithm is to traverse all possible input sequences and construct the next state according to the input sequence.According to the two parameters of the input sequence and the state,the path that meets the requirements in the convolution code tree graph is found out,the input⁃state⁃output transfer table of the convolution code is identified,and then the equation set is constructed to estimate the generation matrix.The experimental results show that the proposed algorithm can effectively perform the blind recognition of the convolutional codes,and avoid the multiple possible generation matrices in the process of backstepping the generation matrices from the check matrix.