Performance of an Adaptive Notch Filter for Spectral Analysis of Coherent Lidar Signals

Abstract

An adaptive notch filter (ANF) is proposed for range-resolved frequency estimates of Doppler lidar atmospheric returns. The ANF is based on the spectral filtering of lidar return to remove the atmospheric contribution from noise. An adaptive algorithm is used to retrieve the filter parameters at a time k knowing both the input signal and filter output at times k ? i, where i = [1, k]. It is shown that ANF performs well at low SNR (?5 dB) compared to the poly-pulse-pair (PPP) estimator currently used for Doppler lidar signal processing. The standard deviation of frequency estimates is 0.01Fs ? 0.02Fs (Fs is the sampling frequency) at SNR = ?5 dB, depending on the signal spectral width. It corresponds to a wind velocity uncertainty of 2?4 m s?1 for Fs = 40 MHz and a laser wavelength ? = 10 µm. The ANF also proved to perform better than PPP in tracking a time-varying frequency, and in the presence of a colored noise.