将噪声对消理论应用于同类混合信号中微弱信号的提取，分析了信号相关性对提取结果的影响。在Simulink中仿真了相关2FSK信号使用噪声对消理论提取期望信号的误码率。对原有的噪声对消工作模式做出改进，提出了基于训练模式的工作方式。仿真结果表明，在新的工作模式下，相关的2FSK信号能够很好地分离，系统在高斯噪声条件下的误码率曲线和理论曲线基本重合。

In this paper the noise cancellation theory was applied to the extraction of weak signal in similar mixed signal, and the effect of the signal correlation was analyzed. Then the error rate of the 2FSK signal was extracted by using the noise cancellation theory in Simulink. To further improve the original noise cancellation work mode, a novel framework based on training mode was proposed. The simulation results show that the 2FSK signal can be well separated under the new operating mode, and the error rate curve and the theoretical curve of the system under the condition of Gauss noise are basically coincident.