采用改进的FastICA算法解决铁路通信同频干扰问题探究  

Research on Improved FastICA Algorithm in Solving Railway Communication Same Frequency Interference Problem

作　　者：刘勇 LIU Yong(China Railway 25th Bureau Group Corporation Limited,Guangzhou 510600,China)

机构地区：[1]中铁二十五局集团有限公司,广东广州510600

出　　处：《现代信息科技》2022年第9期79-83,共5页Modern Information Technology

摘　　要：在复杂的无线电环境下,无线通信和铁路车地通信普遍存在同频干扰,针对在时域和频域内难以解决多个同频信号干扰的问题,文章将盲分离技术应用于同频信号的提取中,提出一种基于盲源分离提取同频信号的方法。首先通过共轭梯度法对基于峭度目标函数的FastICA算法进行迭代,然后在共轭梯度法迭代的基础上增加一维精确线性搜索,最后选用不同代价函数的梯度对步长进行优化,进一步提高了FastICA算法的提取精度。仿真结果表明,所提出的算法能够很好地实现同频信号的提取,提取信号与原信号的相似度达到97%,该研究为解决铁路通信系统中同频信号的混叠问题提供一个新的解决途径。In the complex radio environment, the same frequency interference is common in wireless communication and railway vehicle ground communication. In view of the problem that it is difficult to solve the interference of multiple same frequency signals in time domain and frequency domain, this paper applies blind separation technology to the extraction of same frequency signals, and proposes a method of extracting same frequency signals based on blind source separation. Firstly, the FastICA algorithm based on kurtosis objective function is iterated by Conjugate Gradient method, and then one-dimensional accurate linear search is added on the basis of Conjugate Gradient method iteration. Finally, the gradient of different cost functions is selected to optimize the step size, which further improves the extraction accuracy of FastICA algorithm. The simulation results show that the proposed algorithm can extract the same frequency signals well, and the similarity between the extracted signal and the original signal reaches 97%. This research provides a new way to solve the aliasing problem of the same frequency signals in the railway communication system.

关 键 词：铁路移动通信 同频信号 FASTICA 代价函数 

分 类 号：TN929.5[电子电信—通信与信息系统]