System identification of a miniature helicopter
Conference Paper
Overview
Identity
View All
Overview
abstract
Micro air vehicles are typically designed for mission profiles including surveillance and reconnaissance, and are envisioned to have a large degree of autonomy. Helicopters provide a useful vehicle design as they may carry visual sensors, maneuver through cluttered environments, and hover. Knowledge of the vehicle dynamics facilitates the use of model-based state estimation and control techniques, which can be used to improve sensor measurements, augment pilot handling qualities, and improve autonomous flight performance. Towards that goal this work presents the identification of a linear model for a miniature electric helicopter in hovering flight. The model structure is built upon first principle modeling, previous work, and the statistical contribution of candidate regressors to the model accuracy. Parameter estimates and error bounds are estimated using maximum likelihood methods in both the time and frequency domains, and resulting models are validated by comparing simulated outputs to measured flight data. Results show that the identified models have a Eigenstructure consistent with and predictive capabilities similar to a previously identified model which employed the frequency response method. Copyright 2008 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved.
