3 Day Open Conference

Filtering and Statistics of Flight Test Instrumentation Data
Tuesday 18th November 2008
14.30 - 15.00
Auditorium 3

The US Air Force Test Pilot School relies on flight test instrumentation data to validate aircraft performance and to test systems to their design specifications. Recorded data are sampled at relatively high rates and contain unavoidable noise, both atmospheric noise within air data measurements and vibration noise within accelerometer data. A pitfall of some data analysis techniques is to initially decimate the data to a more manageablesample rate; however, the consequence is a loss of information thus altering the statistically significance of any resulting data analysis. Alternatively, filtering is often applied to a data stream to remove high frequency noise, typically resulting in an artificially high confidence in the results. This paper intends to explore the impact of decimation and filtering on confidence
intervals and to analyze multiple data reduction methods to accurately capture the relative data content, as supported by statistical theory. Monte Carlo simulations will be performed to provide an initial assessment of the performance of the methods under consideration. Results will include analysis of actual flight test data to validate the proposed methods and used to illustrate the consequences of improper data analysis techniques.

About the Speaker(s):

Elwood T. Waddell, Jr., Timothy R. Jorris, Anthony M. Quirarte, Matthew S. Layman, David M. Ho