Research On Modulation Recognition For Signals In Satellite-links

The way from satellites is the most important one to win information superiority inmodern war. It has been a typical characteristic of information-based battlefield toacquire，transmit and distribute various informations through satellites. Therefore，it isa vital path for wining space information superiority to research on the SatellitesInformation-Link Countermeasure.This thesis researches on key basic techniques for sorting and recognizing thesignals in satellites information-links. Satellite Links includes the observe-control link，communication link and data transfer link, which have different characteristics.Through analyzing some typical satellites systems, we have realized that theobserve-control signals are often of multi-mode modulation, the communicationsignals are of either analogue or digital modulation, and the data transfer signals areusually of multi-phase-modulation.Furthermore，the thesis indicates that the commonly used satellite signals involvessuch forms as PCM-BPSK-PM，PCM-2FSK-PM，AM，PM，CW，ASK，DSB，FSK，MSK，BPSK，QPSK，π/4QPSK，8PSK and DSSS，which should be referredin our research. For these signal forms，we have analyzed their characteristics，builtmathematics models and performed a lot of simulations in time and frequency domainwith MATLAB software.Satellites signal modulation recognizing methods can be classified into two kinds，i.e. the maximum likelihood estimation ones and the feature-extraction-based patternrecognition ones. Through comparing spectrum analysis with high-order cumulant andspectral correlation，we confirm the modulation recognition way based on spectrumanalysis. The recognition is based on spectral features，as well as the classificationfeatures which are of better robustness to S/N and modulation parameters. For thequadric-modulation signal recognition，we use the demodulation-recognition schemefor inner-modulate. Besides，the commonly used method of frequency estimating isintroduced. Some related simulation and comparison are completed by using themethods of spectrum peak value position technique，multi-correlation and CZTalgorithm. The frequency estimation of quadric-modulation signal is also discussed.The CZT algorithm is employed as frequency estimating method. Aiming at the PNcode detection in low signal/noise ratio(SNR) signals in direct sequence spectrum spreading(DSSS)，the thesis presents an improved method based on quadric-powerspectrum processing.A prototype based on intermediate-frequency processing is designed for sorting andrecognizing satellite signals. According to the requirements in a project，1GHz highspeed A/D converter module is developed，which is capable of processing signals of300M bandwidth. The signal processing board is based on the design idea of universalhardware platform with DSP+FPGA，so that the prototype is characterized by easyupgrading，easy extending and modularized configuration. The software interface isdesigned with BCGControl Bar，to be compatible with windows XP system.The simulation of satellite signals are generated with vector signal generator andsome self-made devices，while the SNR of the signals can be changed by changing theoutput intensity of a white Gaussian noise generator. The SNR of the signals for themodulation recognition and accurate frequency measure is set at-3dB，while that forparameter extraction of the DSSS is set at-10dB. The recognition correct probabilitycan be higher than80%and the frequency measure accuracy can reach to500Hz.