|
doi:10.3850/978-981-08-7723-1_P158
Neural Networks based System Identification for an Unmanned Helicopter System
Syariful Syafiq Shamsudin1, Xiaoqi Chen2,a, Wenhui Wang2,b, Christopher E. Hann2,c and Geoffrey Chase2,d
1Faculty of Mechanical and Manufacturing Engineering, Universiti Tun Hussein Onn Malaysia, 86400 Parit Raja, Batu Pahat, Johor, Malaysia.
syariful.shamsudin@pg.canterbury.ac.nz
2Mechanical Engineering Department, University of Canterbury, Christchurch 8041, New Zealand.
axiaoqi.chen@canterbury.ac.nz
bwenhui.wang@canterbury.ac.nz
cchris.hann@canterbury.ac.nz
dgeoff.chase@canterbury.ac.nz
ABSTRACT
This paper presents a system identification method to model the dynamic of a small scale helicopter model using neural network model. In order to design an effective autopilot system for an Unmanned Aerial System (UAS) helicopter system, the dynamics of the vehicle platform need to be sufficiently modeled accurately. A method for system identification using neural networks was developed where the test data was provided from a nonlinear dynamics simulator. The acquired test data was then used to train a neural network ARX (AutoRegressive structure with eXtra inputs) model to predict helicopter dynamics nonlinear response. Validation results indicated that the predictions from the training process produce good accuracy and reliable prediction for different set of data.
Keywords: Neural network, System identification, Nonlinear model, Helicopter dynamics, UAS.
Back to TOC
FULL TEXT(PDF)
|