Analysis of weighted subspace fitting and subspace-based eigenvector techniques for frequency estimation for the coherent Doppler lidar.

Since the periodogram maximum (PM) algorithm fails to provide consistent estimates, more robust techniques are developed, especially in a low signal-to-noise ratio (SNR) regime. The methods are formulated in a subspace fitting-based framework, such as the eigenvector (EV) method and the proposed weighted subspace fitting (WSF) method by introducing an optimal weighting matrix, which exploits the low-rank properties of the covariance matrix of the coherent Doppler lidar echo data. Simulation results reveal that the number of the reliable estimates by the WSF method is more than the other two methods, and the standard deviation is the smallest. Furthermore, the predicted best-fit Gaussian model for the probability density function of the estimates has a narrower spectral width than that of PM and EV methods. Experimental results also validate the simulation results, which show that the WSF approach outperforms the PM and EV algorithms in the furthest detectable range. The proposed method improves the detection range approximately up to 14.2% and 26.6% when compared to the EV method and the PM method, respectively. In conclusion, the proposed method can reduce the statistical uncertainties and enhance the accuracy in wind estimation specifically for a low SNR regime.