Abstract:

To estimate the parameters of linear frequency modulation (LFM) signal precisely in a noisy  environment, a novel discrete spectrum correction method is proposed based on Radon-ambiguity transform (RAT) and fractional Fourier transform (FRFT). This method reduces the parameter estimation to two one-dimensional spectral peak searching processes, and by the spectral analysis of RAT and FRFT of LFM signal multiplied by rectangular window, the energy centrobaric correction measure is introduced in these two searching processes, which improves the estimation accuracy substantially without increasing any computing complexity. As shown in the simulation results, for the singlecomponent LFM signal case, the correction errors of the RAT and FRFT spectral peak locations are reduced to 1.55% and 4.94% of the scanning step, which verifies the efficiency of the method.