中文摘要：

针对现有的矩阵分析法对线性分组码进行盲识别时,容错性能较差的问题,提出了一种改进的方法。首先利用截获的码字数据建立分析矩阵并进行高斯消元,然后计算各列列重的归一化值,按照判决门限找出分析矩阵中的线性相关列,并以此建立统计量,最后通过统计量极大值的分布规律完成码长的识别。识别出码长后,通过移位处理及随机交换分析矩阵的行进行多次平均,实现高误码率下码字起点的识别。仿真结果表明,该方法与传统矩阵分析法相比,计算量基本相当,但容错性能有很大提升,能在较高误码率下有效实现线性分组码的盲识别。

英文摘要：

Since the existing matrix analysis method has a poor performance when solving the blind recognition problems of linear block codes, an improved method is proposed. Firstly, Gauss elimination algorithm is applied to the interception matrix and the normalized weight of each column is calculated, then a statistical esti- mation is used to find correlation columns in the matrix, based on which a statistical index is setted up to get the code length according to the distribution of maximum values. To achieve successful recognition of code start bit under high bit error rate （BER）, shift operation with many times of average is adopted after code length identifi- cation. Simulation results show that the proposed method has a better fault-tolerant performance when comparing to the traditional matrix analysis method, while the computational complexity is approximately the same, thus it can be more effective for the blind recognition problem.