Study Of Track Signal Demodulation Method Based On TMS320C6720

With the rising speed of the train, how to receive much faster and more accurately track signal has becomes the key problem of raising speed and improving operative efficiency. At present, in vehicle system aspect, it is small spase for railway trains to raise speed because of slow operational speed of the main processor; And in frequency detection aspect, the detection quality is downtrend because of raising train speed. In this paper, aiming at these problems, it puts forward an improved scheme, (1) in system aspect, upgrading the vethicle system, with the much faster processors for demodulation to replace the original processors; (2) in algorithm aspect, developing a new more stable and reliable demodulation algorithms to enhance the detection quality and detection means.In this paper, at the TMS320C6720 hardware platform, it analogs the railway field signals and implements some common demodulation algorithms currently used based it, including spectrum analysis, IFM by frequency deduction and IFM by phase deduction. the TMS320C6720 hardware platform has the faster operation speed and the higher demodulation precision, and fully embodies the real-time of demodulation compared with the hardware platform of the TMS320VC33. So it is completely feasible to use chip C6720 to replace the will soon stop production chip VC33.In algorithm aspect, at present, the railway signal demodulation method is using IFM by the frequency deduction or IFM by phase deduction to calculate the upper and lower side frequencies of rail-signals, and using the spectrum analysis to demodulate the modulation frequency. The shortcoming of this method is that it can meet the precision requirement, only when signal-to-noise ratio is very high. In this paper, based on this, it researches a new demodulation scheme that the Hilbert-Huang Transform (for short HHT) is introduced into the railway FSK signal demodulation. And based on the HHT, it studies out improved algorithms. And using this improved algorithm, it only needs to use the instantaneous frequency of IMF1 (the first Intrinsic mode function) during demodulating the upper and lower side frequencies, carrier frequency and modulation frequency of the railway FSK signal. The simulation results show this method's demodulation precision is higher than the three methods mentioned above, even if when the in-band (450Hz-2700Hz) signal-to-noise ratio is 1:3, or even lower. Especially for the low frequency, it is two orders of magnitude more than the frequency domain analysis. So this improved algorithm can be used field practical environment.