A Method of Noise Reduction for Chirp Signal Based on K-L Orthogonal Division

Because the estimation of instantaneous frequency (IF) can be easily affected by noise, a method of noise reduction for Chirp signal based on K-L orthogonal division is presented. The signal is concentrated to a principal component, but the noise is equally distributed to several components. Thus, the SNR is improved. Both analysis and simulation indicate that this method works well and makes the estimation of IF more accurate. And the SNR can be improved over 6 dB. Moreover, associated with the characteristic of Chirp's auto-correlation function, the simplification of this method is considered for some applications.