Time Domain Identification Of A Helicopter Using Minimal Representation And Time Delays

Abstract

This paper presents a standalone time domain approach to identify the dynamics of a full-scale helicopter in both hover and forward flight. A high-fidelity nonlinear helicopter model of a commercial 3t classd helicopter is employed to generate the dataset, which consists of 3211 and 2311 type inputs for identification and frequency sweep type maneuvers for verification. The output error algorithm is used in conjunction with the Levenberg Marquardt routine in time domain to estimate the parameters of the state space model structure. A systematic model reduction routine is employed to eliminate redundant parameters from the model struc- ture, thereby achieving a minimal representation of the dynamics. To account for the higher-order dynamics induced by the rotor, time delays are introduced into the state and control matrices in the relevant deriva- tives. The identified models are verified in frequency domain by demonstrating that the errors fall within the acceptable bounds. Furthermore, the obtained set of parameters is shown to be consistent with the values reported in the literature.