Parameter estimation of non-modulated or modulated Frequency-Hopping signals

It is a very effective way to use time-frequency distribution to analyze the Frequency-Hopping (FH) signals. There are a variety of time-frequency analysis methods, in which wavelet transform's time-frequency distribution of the signal is very sensitive to noise, and Wigner-Ville distribution has a good time-frequency aggregation but the presence of severe crosstalk analysis of multi-component signals. Classic STFT is a good time-frequency tools, but cannot obtain a higher time resolution and frequency resolution at the same time. In this paper, the classical STFT algorithm is improved to work well in lower SNR by using image processing, and further improve time resolution at the same time combined with the differential frequency discrimination in high SNR. Experimental results show that reasonable input parameters will improve the performance of frequency hopping signal parameter estimation.