Abstract

Coherent Doppler lidar has become a mature method of remote measurements of wind fields[l, 2, 3]. The development of the eyesafe 2µm coherent Doppler lidar has permitted regular observations of boundary layer winds. Coherent Doppler lidar permits measurements of the radial component of the velocity v as a function of range for every transmitted pulse. As the target range increases, the signal level decreases until the quality of the estimates degrade due to many random outliers. The quality of the velocity estimates can be improved by accumulating the signal from many consecutive pulses[4, 5, 6] provided the wind field is not changing over the accumulation time. Rye and Hardesty investigated the effects of pulse accumulation by comparing the standard deviation of the velocity estimates to Levin’s approximation to the Cramer Rao Bound (CRB), the theoretical best performance. It has been shown that Levin’s approximation to the exact CRB is a good approximation[7]. For very weak signals, the pulse accumulation is insufficient and the standard deviation is dominated by the random outliers. As the number of accumulated pulses increases, the random outliers decrease and the standard deviation approaches the CRB. Rye and Hardesty defined n_CR as the number of accumulated pulses that were required to first approach the CRB. An approximate expression for n_CR was produced, i.e., n_CRM = 64/SNR^3/4 where M is the number of complex data points per estimate and SNR is the wide band Signal to Noise Ratio for one shot.

© 1995 Optical Society of America