Modulation Recognition And Parameter Estimation Of Wideband Frequency Hopping Signal

Frequency hopping is an important way to realize spread spectrum communication with high spectrum utilization,Frequency hopping have outstanding anti-interference and anti-acquisition capabilities because its signal carrier frequency hops randomly,so it plays an important role in military and satellite communication.As the basis for the next demodulation module,parameters estimation and modulation identification of frequency hopping signal will be the focus of research.The actual frequency hopping system has the characteristics of high speed frequency hopping,short hopping cycle,less data,the filter needs to be adaptive during signal processing owing to indefinite bandwidth of each signal,so it is difficult to analyze each signal.Under the above frequency hopping system,this thesis mainly focuses on the bandwidth,carrier frequency and symbol rate estimation of DPSK,SDPSK and GMSK signals in frequency hopping system,then identifies the modulation mode of signal.In this thesis,the Welch method is used instead of periodogram method to estimate the 3dB bandwidth of signal.Simulation results indicate that Welch method can significantly improve the variance characteristics of the spectral estimation.In consideration of each signals' characteristics,the relationship between the signal squared spectrum and the carrier frequency and symbol rate can be obtained.Carrier frequency estimation of all this three modulation signals can use the same algorithm.A rough carrier frequency is estimated by center frequency method,then using square spectrum method to estimate carrier frequency accurately.In order to further improve the estimation accuracy and reduce the computational complexity,CZT algorithm is used to refine the spectrum after the square spectrum estimation.As a result,compared with square spectrum algorithm,square spectrum and CZT algorithm reduces the estimation error of carrier frequency from about 0.1% to less than 0.01%,further more,computational complexity is reduced by nearly 3 times.Since the square of SDPSK and GMSK signals is periodic,and its period is related to the symbol period,square spectrum and CZT algorithm can be used to continue the symbol rate estimation of two signals.The period of the envelope of the DPSK signal is a symbol period,so the symbol rate of the DPSK signal can be estimated by amplitude spectrum method.The simulation results show that above two estimation methods have high precision for the symbol rate estimation of each signal.From the square spectral characteristics of each signal,DPSK and SDPSK,GMSK signals can be identified by extracting the number of spectral line of the square spectrum.This thesis presents a method to extract the flatness coefficient between two strong spectral line.When SNR is 12 dB,after 2000 simulations,the correct recognition rate for three signals is over 99%,at the same time,it can maintain a high recognition rate at low SNR.In the actual project,the algorithm introduced in this thesis can accurately and effectively realize the modulation identification and parameter estimation of frequency hopping signal.