Online parameter estimation of a miniature unmanned helicopter using neural network techniques
Date
2012-02-01
Authors
Kantue, Paulin
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The online aerodynamic parameter estimation of a miniature unmanned helicopter using
Neural Network techniques has been presented. The simulation model for the miniature
helicopter was developed using the MATLAB/ SIMULINK software tool. Three trim conditions
were analyzed: hover flight, 10m/s forward flight and 20m/s forward flight. Radial
Basis Function (RBF) online learning was achieved using a moving window algorithm which
generated an input-output data set at each time step. RBF network online identification was
achieved with good robustness to noise for all flight conditions. However, the presence of
atmospheric turbulence and sensor noise had an adverse effect on network size and memory
usage. The Delta Method (DM) and the Modified Delta Method (MDM) was investigated
for the NN-based online estimation of aerodynamic parameters. An increasing number high
confidence estimated parameters could be extracted using the MDM as the helicopter transitioned
from hover to forward flight.