| Frequency hopping communications is one of the main forms of non-cooperative communication because of its carrier frequency changes according to some pseudo-random sequence. It difficult for third parties to intercept without any prior knowledge of the case, so it has good privacy and has been widely used. However, many times, we need to extract the frequency-hopping signal parameters without any prior knowledge, thereby obtaining information. This is a very large value to national security. In this paper, a comprehensive and in-depth research is taken about hopping signals. In order to analyze the macro and micro various parameters of frequency hopping signals, and thus realize the de-hopping and demodulation, and ultimately to extract information, and to restore the ideal modulation signals through information extracted. And then analysis of the modulation quality of the signal. Hopping signal analysis system has been established. And is implemented in software.Information extraction of hopping signal is divided into de-hopping and demodulation. We have studied two solutions about de-hopping. One is hopping pulse analysis methods to analyze the signal power. This method can achieve the split pulse waveform using automatic adaptive segmentation, enabling the de-hopping. And then through detailed analysis of the pulse, the parameters of transition time, pulse width, pulse overshoot are estimated. Classic algorithms are given. Experiments show that the algorithm can quickly and effectively split hopping pulse waveform accurately estimate the various parameters of frequency hopping pulse signal, to achieve demodulation of frequency hopping signals.Another solution of de-hopping is to analyze the frequency-hopping signal hopping pattern through frequency analysis method, which is the classic method of analysis of frequency hopping signals. This paper studied the common analysis methods of the frequency hopping signals including short-time Fourier transform frequency analysis method, Gabor transform and Wigner distribution, the pseudo-Wigner distribution, smoothed pseudo Wigner distribution, frequency spectrum analysis methods. Advantages and disadvantages of these methods is studied, as well as their applications. The system selects the classic short-time Fourier analysis, because it has a lower computational complexity. After getting hopping pattern, the filter can be used to select a different frequency channels to extract the frequency-hopping signals. This paper presents a classic FIR design method in order to achieve the de-hopping.By de-hopping the signal, we get the fixed-frequency communication signals. The next step is to analysis parameters of the signal and demodulation, such as frequency offset, symbol rate, timing, signal to noise ratio, bandwidth, phase noise, stray estimate. This paper focuses on the estimation method of symbol rate and the EVM. For symbol estimates, this paper proposes a new MPSK signal symbol rate estimation method based on Haar wavelet,The experiment proved that the method can better estimate the symbol rates in the cases signal symbol number less than10. It is especially suitable for frequency hopping signal symbol rate estimation, which is the other symbol rate estimation methods can not reach. When the data is obtained from the signal, we re-modulate the data to get ideal modulated signal, which is called the reference signal. Calculating the ratio of rms of reference signal and the actual received signal, we get a significant response signal modulation quality Parameters-error vector magnitude (EVM:error vector magnitude). Since it’s a non-collaborative system, symbol rate estimation error will have an impact on the EVM, The article presents a wavelet-based reference signal recovery method, this method can in symbol rate under certain error conditions recover more accurate reference signal, to get a more accurate EVM parameters. Experiments show that the method is robust, suitable for EVM analysis of large amounts of hopping signal data. |
PDF Full Text Request