TIME FREQUENCY ANALYSIS - AN APPLICATION TO FMCW RADARS
Balaji Nagarajan · 2004
The Fourier transform of a signal defines the frequency domain representation of the signal in that it specifies relative amplitudes of the various frequency components of the signal. However, the Fourier transform is not always the best tool to analyze ‘real-time signals ’ which have frequency components that change over time. Joint time-frequency techniques were developed for characterizing the timevarying frequency content of the signal. This project presents an overview of the basic concepts and well-tested algorithms for joint time-frequency analysis with particular reference to their application to radar signals. The time-frequency techniques can be classified into two types: linear (e.g., short-time Fourier transform, Gabor expansion, Wavelet transform) and quadratic (e.g., Wigner-Ville distribution, Cohen’s class of distributions, Timefrequency distribution series). The aforementioned techniques are discussed in detail and are tested first against ideal simulations of normal cosine and chirp signals and then with the radar beat signal. A brief description of FMCW sea-ice radar developed