An Improved Algorithm Based On S-Transform For Parameter Estimation Of Frequency-Hopping Signals

The S-transform can't both have the time and frequency resolution at the same time. So in order to solve this problem, this paper proposes a new joint algorithm for parameter estimation of frequency-hopping signals. Frist, S-transform with short window is obtained. Then, accurate estimation of hopping time is extracted from the time-frequency ridge of S-transform. Finally, using the FFT (Fast Fourier Transformation) algorithm, we can estimate the frequency of the signal between the two hopping times. Considering the distortion of the ridge line at high frequency component, and combining with the statistical property of white Gaussian noise in S-spectrum, an indirect band-pass filtering method in time-frequency plane is proposed to smooth the ridge line and decline the influence of noise. Theoretical analysis and simulation results verify of the effectiveness and feasibility.