Nonlinear Parameter Estimation from Flight Test Data Using Minimum Search Methods

摘要

The maximum likelihood (ML) method, based on minimum search algorithms, for extraction of the stability and control derivatives from flight test data is discussed. General nonlinear multivariable model structure for equations of motion incorporating the nonlinearities due to state and control variables, as well as due to the parameters being estimated, are investigated. Comparisons with the quasi-linearization method and other numerical computational aspects of integration, and accuracy of parameter estimates are presented. Use of ML techniques for processing large amounts of flight test data in a routine manner remains restrictive, due to their slow convergence. Poor rate of convergence for the gradient as well as nongradient methods is irrespective of the number of parameters being estimated.