Abstract: Global System for Mobile Communications for Railway （GSMR） communication system of highspeed railway is vulnerable to the interference of wireless signals of other types. In this paper, the authors analyzed and researched several GSMR interference signals and proposed a new kind of signal separation algorithm based on the multiframe statistical spectrum. The idea of the algorithm was that the interference signal was divided into multiple frames and accumulated to statistical spectrum. If the number of frames was sufficient, its spectral parameter tended to be a constant. The experimental results show that the algorithm can quickly separate the original and interference signals in the case of known original signal spectrum, its effect and versatility is better, and the complexity of the algorithm is less than the traditional method.

Key words: Gaussian Minimum frequencyShift Keying （GMSK）, interference, demodulation, signal separation

摘要： 高速铁路铁路专用移动通信网 (GSMR)通信系统易受到其他类型无线信号干扰。对常见的几种GSMR干扰信号进行了分析和研究，提出了一种基于多帧统计谱的信号分离算法，将被干扰信号谱分成多帧并进行累加统计，当统计帧数足够多时，其频谱参数趋于恒定。实验结果证明，该算法在已知原始信号频谱的情况下可以快速分离原始信号和干扰信号，效果良好且具有通用性，算法复杂度低于传统方法。

关键词: 高斯最小频移键控, 干扰, 解调, 信号分离