Determination of the Stability and Control Derivatives for the Variable-Response Research Aircraft Using a Modified Maximum Likelihood Estimator.

摘要

A maximum likelihood estimation program was applied to flight data for Princeton's Variable-Response Research Aircraft to determine its primary stability and control derivatives. The control derivatives for the side-force surfaces and the rudder were of special interest. The effects of measurement noise and process noise on parameter identification also were studied. This investigation showed that the maximum likelihood estimation program used identifies derivatives which produce close fits of the measured time histories. Standard deviations of the derivations computed from several time histories indicate the quality of the estimates. The reduction in standard deviations when estimates were separated by type and direction of control input time history used indicates that derivative estimates are affected by the assumptions inherent in the analytical model and the signal-to-noise ratios of the data. The method used for identifying highly correlated derivatives also affected the estimates obtained. The final set of derivatives determined in this research produced a good fit of the measured data and several of the derivatives agreed well with analog matching derivative estimates. (Author)