Parameter Estimation for Aircraft Flying in Turbulence

A time varying filter is applied to estimate aerodynamic parameters from aircraft data in turbulence. The filter, employed for state estimation, forms an integral part of the iterative Gauss-Newton method used for estimation of aircraft longitudinal stability and control derivatives through optimisation of the likelihood function. A non-linear model postulate, in conjunction with the Dryden model, is used to simulate aircraft response in mild, moderate and severe turbulence. The simulated data is analysed using the output error and the filter error13; approach. Comparison of the measured and estimated aircraft response time histories is presented. Some other aspects like the estimation of initial conditions, Cramer-Rao lower bounds for the estimated parameters, number of iterations required for convergence and CPU requirements, are also provided. The estimates from filter error method are found to be better than those from OEM in presence of turbulence.