Feature Frequencies Extraction From The Track Frequency Shift Keying Signals

Train running monitor and record devices are the core operating system on-board train, which achieve safe speed control and provide the information of operational status. Their quality will have a direct impact on train punctuality and safety rate. Among them, track signal processing and identification validity of the railway insulation nodes decide the railway transport capacity and operational safety. This paper bases on looking for the relevant literature and analyzing the features of the track circuit frequency-shift signal and requirements of train operation, for the following work.1. This paper analyzed the track circuit and signal features. Based on their features, the track circuit frequency-shift keying (FSK) signal was analyzed with both the time-domain and frequency-domain. The FFT and wavelet transform, which were used to describe the FSK signal, were given.2. The host of existing LKJ2000 train running monitor and record devices were taken as the plant and the work principles of all kinds of the important plug-in were introduced in detail. The status and roles of the ground plug-in message board in the host were discussed. Some defects of the information processing function in the existing plug-in were clarified.3. Based on the ground plug-in message board and the fully consideration for features in FSK, DSP chip with model TMS320C6713 was employed to upgrade the function of processing information on the plug-in. Taking TMS320C6713 as a platform, the simulations and debuggings for FSK were implemented with spectrum analysis and wavelet transform.4. Noticing the features of FSK and its actual differences in the initial conditions, wavelet singularity detection and wavelet ridge were used to extract the features for both the ideal track circuit FSK and the track circuit FSK from LKJ2000 Monitoring Test. These two methods can effectively demodulate the track circuit FSK.Three algorithms were validated in the laboratory. The experiment showed that every algorithm was usable and the frequency errors of the algorithm of wavelet ridge were given. Three algorithms can all demodulate the track circuit FSK and low frequency control signal. By comparing with the existing method with time-domain analysis three algorithms discussed in this paper are predominant in qualities. The algorithms can give effectively judgment for the track circuit FSK and low frequency control signal.