Performance of 10.6 mu m CO2 Doppler lidar at low laser frequency stability using fixed hard target

Performance of CO2 Doppler Lidar for fixed hard target at low laser frequency stability (in the order of 10(-7)) is presented using the novel wide-band Doppler detection technique described before. The approximate maximum likelihood estimator is used. The velocity estimation error is experimentally determined at various SNR's and is compared to the Cramer-Rao Bound. Our results show a deviation of the unbiased velocity error from the CRB less then the width of the confidence interval for the error. The biasing of the estimator at low SNR's (or the presence of so called ''bad'' estimates) is studied experimentally for non-ideal equipment performance (correlated noise); we propose the Gaussian model of the distribution of so called ''bad'' estimates as more appropriate for experimental data, instead of the uniform one, recommended in Ref.5.