Research on Track Frequency Shift Signal Detection Based on Local Mean Decomposition

The track frequency shift signal reflects the status of signal light in front of the train. Due to the development of high-speed railways, the main demodulation method, improved Fast Fourier Transform (FFT) is unable to meet requirement of limited sample time. In this paper, a new signal demodulation method, Local Mean Decomposition(LMD) is introduced to detect frequencies of the signal. LMD adaptively decompose complicated signal into a set of single-component signals, each of which has physical meaning. In order to solve end effect of LMD, an adaptive waveform matching extending method is proposed. In addition, a noise-assisted analysis method is recommended to reduce the noise effect. The simulation results show that compared with the improved FFT, LMD has better performance on detecting track frequency shift signal. LMD nearly has no requirement on signal sampling time but has twice the accuracy of FFT with 5s sampling time.