Adaptive filters for frequency estimate of heterodyne Doppler lidar returns: Recursive implementation and quality control

Unreliable frequency estimates at low signal-to-noise ratios, provided by a heterodyne Doppler lidar (HDL), undermines any data analysis scheme requiring high-accuracy wind fields retrievals. To meet demanding specifications, that is, high accuracy associated with high reliability on radial velocity components, a quality control (QC) procedure has to be implemented at signal processing level. The authors propose use of a recursive implementation of an adaptive filter for frequency estimate coupled with a QC that combines a statistical test on the signal energy filtered out as proposed by Rye and Hardesty and a persistency criterion (PC). The PC leads to improved performance with respect to the percentage of data finally accepted and frequency accuracy. The performance is validated using simulated signals and HDL actual data.