Modelling and identification of nonlinear aerodynamic functions using B-splines

A method of using B-splines to represent nonlinear aerodynamic functions in flight dynamic modelling is presented. This method offers advantages over traditional lookup table and polynomial representations for the problem of parameter identification from flight test data. The method offers a natural fit to an underlying smooth function with guaranteed continuity properties. In addition, the B-spline coefficients offer improved conditioning to the identification process, promising improved convergence properties of identification algorithms. A reduction in the number of flight tests required is possible, since large portions of nonlinear functions can be identified from large-amplitude manoeuvres. In addition, when used with the maximum likelihood estimation method, the Cramer-Rao estimates of parameter confidence are preserved and can be expressed in a B-spline format