Optimal Wind Velocity Estimation.
ARMY ELECTRONICS COMMAND FORT MONMOUTH N J
Pagination or Media Count:
A new approach to remote wind velocity sensing is proposed, developed, and tested. Wind estimates were obtained from the spectral signature of an optical signal source. The signal was derived from the random scattering of a laser beam in the lower atmosphere. A quasi-supervised learning machine, using a minimal variance linear estimator, was used to produce the wind velocity estimate. New stopping rules, used to guarantee the convergence of the learning process, were also developed. Field experimentation indicates that the new system will perform well even under extreme optical nonlinear conditions severe turbulence.