序列检测和近似熵检测的快速实现研究

随机序列广泛应用于信息安全领域，随机序列的质量依靠随机性检测规范判断。美国国家标准与技术研究院提出的随机性检测规范中包含序列检测和近似熵检测，这两种检测算法的运算速度位列检测包末端。对两种检测算法的运算流程进行研究，通过优化字节处理方式、字节运算与相对频数统计相结合、频数统计值复用等方法分别优化两种算法，并将两种检测算法进行合并，减少冗余的数据加载和处理流程，完成两种算法的快速实现。实验结果表明，不同参数下序列检测和近似熵检测速度最高分别提升30.02倍、27.58倍，两种检测合并后整体速度最高提升45.23倍。

Study on Fast Realization of Serial Test and Approximate Entropy Test

Random sequences are widely used in the field of information security, and the quality of random sequences depends on randomness test specifications. Randomness test specification proposed by the National Institute of Standards and Technology of the United States includes serial test and approximate entropy test. The operation speeds of the two tests algorithms are listed at the end of the detection packet. The operation flow of the two tests algorithms is studied. By optimizing byte processing mode, combining byte operation with relative frequency statistics, and multiplexing frequency statistics, two tests algorithms are optimized respectively. The two detection algorithms are merged to reduce redundant data loading and processing flow, and the fast implementation of the two algorithms is completed. The experimental results show that the maximum speed of serial test and approximate entropy test increases by 30.02 times and 27.58 times respectively under different parameters, and the maximum overall speed increases by 45.23 times after the combination of the two detection methods.