Two-Stage Estimator for Frequency Rate and Initial Frequency in LFM Signal Using Linear Prediction Approach

We propose a two-stage estimator to estimate chirp rate and initial frequency of the chirp signals in the presence of additive white Gaussian noise. In the first stage, the chirp rate estimation problem is reformulated as a single tone frequency estimation problem. Then, the frequency of single tone is estimated through a linear prediction approach. Using the chirp rate estimate in the first stage, we can convert the linear frequency modulated signal to a single tone. Similar to the first stage, the initial frequency is estimated via the linear prediction approach. The performance of the present method is assessed by comparison with Cramer-Rao lower bound and other existing methods through computer simulations. The proposed algorithm estimates well for different values of the chirp rate and initial frequency, as well as for different number of samples. In other words, this algorithm has uniform performance for various values of the signal parameters.