基于时频分析的移频轨道交通信号检测方法  

Frequencyshift Rail Transit Signal Detection Method Based on Time-frequency Analysis

作　　者：潘长玉[1] PAN Changyu(Communication Signal Design Institute of China Railway First Survey and Design Institute Group Co.,Ltd.,Xi’an 710043,China)

机构地区：[1]中铁第一勘察设计院集团有限公司通信信号设计院,陕西西安710043

出　　处：《机械与电子》2022年第1期71-75,共5页Machinery & Electronics

摘　　要：针对传统方法对交通信号检测时,由于未能提取交通信号的时频特征,导致信号检测时存在检测精度低、检测误差大和噪声频率不稳定等问题,提出基于时频分析的移频轨道交通信号检测方法。首先利用时频分析法对移频轨道交通信号的时频特征进行提取;再基于提取的信号时频特征,利用信号的概率密度函数获取交通信号的信号双谱;最后利用卷积神经网络分类处理有双谱的交通信号,实现信号检测。实验结果表明,该方法检测信号时,检测精度高、检测误差小,以及噪声频率稳定。In view of the problems of low detection accuracy,large detection error and unstable noise frequency in the process of traffic signal detection due to the failure of traditional detection methods to extract the time-frequency characteristics of traffic signals,a frequency-shift rail transit signal detection method based on time-frequency analysis is proposed.Firstly,the time-frequency characteristics of frequency-shift rail transit signal are extracted by time-frequency analysis method;Then,based on the extracted time-frequency characteristics of the signal,the traffic signal bispectrum is obtained by using the probability density function of the signal;Finally,the bispectrum of the traffic signalsare classified and processed by convolutional neural network to realize signal detection.The experimental results show that high detection accuracy,small detection error and stable noise frequency can be achieved through this detection method.

关 键 词：时频分析 交通信号 移频轨道 检测方法 卷积神经网络 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]