An Application of Kalman Filtering to Airborne Wind Measurement

Abstract

Airborne wind measurement techniques currently being used onboard the National Aeronautical Establishment (NAE) Twin Otter Atmospheric Research Aircraft are described and their fundamental limitations are discussed. In particular, a recently acquired LTN-90-100 strapdown Inertial Reference System (IRS) exhibits significant low frequency errors in its velocity components (primarily Schuler oscillation errors that can attain peak values of 2 to 3 m s?1), actually degrading wind computation accuracy compared with older techniques. A new wind measurement technique, based on a Kalman filter integrated navigation approach, is shown to mitigate this problem and provide wind computation accuracy superior to previous methods. Preliminary results, based on applying the Kalman filter to Twin Otter flight test data, indicate that inertial velocity accuracies of 0.3 m s?1 rms (per axis) are attainable under ideal conditions, with a corresponding improvement in the accuracy of earth-referenced wind components.