Parameter Estimation And Modulation Recognition For High Speed Frequency Hopping Signal

Since the Frequency-Hopping(FH) communication possess the advantages such as low probability of interception, anti-interference ability, etc., FH technology was widely used in military communications. However, in order to carry out reconnaissance and jamming the FH communications efficiently in the field of communication confrontation, parameter estimation and modulation recognition technology has became a key of the signal processing, and a necessary function of the signal reconnaissance receiver.Firstly we need to conduct Time-Frequency analysis in the FH signal processing. So the research on the Time-Frequency analysis algorithm will be the key content of this thesis. In this paper, Short-time Fourier transform, Gabor transform, Wavelet transform and Time-Frequency Distribution of Cohen are used for Time-Frequency analysis of high-speed FH signal to study the performance of these algorithms respectively, with a lot of simulation experiments. In the Time-Frequency Distribution of Cohen analysis algorithms, Smoothed Pseudo Wigner-Ville Frequency Distribution(SPWVD) does not only has high time resolution and frequency resolution but also has a good Time-Frequency aggregation. But the calculation is very large, which is not suitable for real time processing of FH signals. However, it's found that Short-Time Fourier transform has low frequency resolution and Time-Frequency concentration, but the small amount of calculation is suitable for real time signal processing. Therefore, in order to facilitate the project implementation and complete real time processing of FH signals, we apply Short-Time Fourier transform to perform Time-Frequency analysis of FH signals. We work with the Short-time Fourier transform to estimate the FH moments, combined with the least squares method to estimate the FH period accurately.In FH signal processing, it is easy to get each hop signal by estimating FH moment. In this paper, each hop signal processing contains modulation identification, frequency estimation and symbol rate estimation. For different modulation signals, we propose corresponding algorithms about frequency estimation and symbol rate estimation. For modulation recognition of communication signals, this paper proposes a new algorithm based on fast intermediate frequency signal modulation recognition. Compared with the current conventional modulation recognition algorithm, the new algorithm just needs to extract five feature parameters to complete modulation recognition by the tree classification, and this algorithm doesn't need precise frequency and symbol rate. Finally, we completed the parameter estimation and modulation recognition about FH signal with the DSP-based hardware platform, which shows that the algorithms proposed in this paper can be realized.