David H. Marlin
U.S. Army Res. Lab., Battlefield Environment Directorate, ATTN: AMSRL-BE-S, White Sands Missile Range, NM 88002
Past research has shown that atmospheric turbulence, as opposed to wake shedding and aeroacoustic noise, is frequently the predominant cause of wind noise in microphones. Wind speed measurements from hot-film anemometers generally show good correlation between the turbulent wind speed and the pressure fluctuations as measured by a microphone in close proximity to the anemometer. The purpose is to discuss attempts to experimentally estimate the transfer function between a hot-film anemometer and microphone. Some ``black-box'' input--output techniques, including ARMA, adaptive filtering, and neural networks are considered. Also discussed are state-space estimators such as the Kalman filter, using simple dynamical models of the measurement devices and theoretical models of the pressure-velocity relationship.