Instantaneous Frequency Estimation Based On The S-method

Doppler shift signal is a non-stationary signal, its instantaneous frequency is afunction of time, and changes over time. The traditional Fourier Transform cannot revealthe time domain information and frequency domain information at the same time, andcannot be used to process non-stationary signal. Time Frequency Analysis can effectivelyshow the signal frequency variation law with time, Time Frequency Distribution is thebasis and core, there are a variety of forms of Time Frequency Distribution, Short TimeFourier Transform and Wavelet Transform is the common linear time-frequencydistribution, subject to fixed window function, Short Time Fourier Transform (STFT)cannot be able to obtain high time resolution and high frequency resolutionsimultaneously, it is difficult to design the basis function for one special signal by usingWavelet Transform. Wigner Ville distribution is the simplest quadratic Time-FrequencyDistribution, it has a good auto-terms aggregation, but because of the existence of thecross-terms, it cannot be used for the analysis of the multi-component signal. S-methoddistribution can make a good compromise in the auto-terms aggregation and cross-termssuppressing.This paper presents an improved algorithm based on the S-method distribution,named Modified S-method distribution, it is realized by changing the area of the windowfunction, in stable area, a window with large area is used, in poor stable area, a windowwith small area is chosen. The simulation experiments show that compared with S-method distribution, the Modified S-method distribution has higher auto-termsaggregation, and achieved considerable Instantaneous Frequency Estimation precision.Modified S-method distribution applies to the actual signal, for different types of Dopplersignal analysis show that the Modified S-method has better time-frequency aggregation,can more clearly describe the instantaneous frequency of the signal. Finally, in order tofacilitate signal analysis, a graphical user interface is designed for signal processing.