Application And Research Of Adaptive Filtering Algorithm Of FSK Signal Demodulation In Track Circuit

With the rapid development of railway transportation and the continuous improvementof train running speed, it puts forward higher requirements for the safety of railway. Trackcircuit is widely used in national railway system as a high reliability of railway infrastructure.The precise demodulation of Frequency Shift Keying(FSK) signal as the main transmissionsignal in the track circuit is prerequisite to ensure driving safety. At present the FSK signaldemodulation of track circuit mainly uses spectrum analysis method based on fast Fouriertransform to complete, but the method is not good to low frequency modulation frequency indemodulation accuracy and it will affect the real-time of the signal demodulation to improvethe accuracy. Therefore, it needs to study a reliable, accurate and good anti-interferencedemodulation method to complete the track circuit FSK signal demodulation accurately inreal-time.Based on adaptive filter is not sensitive to the undesired signal, the thesis introduces theadaptive filtering in the track circuit FSK signal demodulation. As the adaptive filteringtechnology has high demodulation precision, good real-time and strong anti-interference tospectrum analysis and it can accurately get the FSK signal carrier frequency information, thethesis establishes the demodulation model by using above advantages to realize the simulationof FSK signal in the demodulation of track circuit.Firstly, the thesis expounds the principle and time-frequency characteristics of trackcircuit FSK signal and analyses the advantages and disadvantages of the existingdemodulation technologies. Through the analysis of the principle of adaptive filteringtechnique, it introduces varying step-size Variable Step-Least Mean Square(VS-LMS)algorithm in view of the defects of traditional Least Mean Square(LMS) algorithm in theiteration step length values, and it gives the contrast analysis of filter performance of them.Secondly, it combines spectrum analysis and adaptive filtering technique to create theFSK signal demodulator model for track circuit. It can obtain input signal carrier frequency ofthe system through the method of spectral analysis and takes one of the side frequency asdesired signal of adaptive filter system. Then it extracts a carrier frequency component torestore the modulation wave, and it completes the signal demodulation after the calculation ofmodulation frequency.Finally, the thesis implements two track circuit FSK signal demodulation respectivelywhich is the domestic frequency shift and ZPW-2000A in the case of Gaussian white noiseand sinusoidal noise by the use of Matlab simulation platform. The simulation results provethat this method has good real-time performance, strong anti-interference ability and highdemodulation accuracy compared with the signal standard in railway signal classification strategy. The method can make up for the inadequacy of existing demodulation technology,and it has certain practical significance to achieve for high-precision real-time track circuitFSK signal demodulation.