Tor-Aleksander Johansen, Annette Fossli Brustad, Tom Stian Andersen, and Raymond Kristiansen
System Identification, NARMAX, FROLS, Linear Regression
We apply the NARMAX approach for nonlinear system identification in order to evaluate it for identifying the aero- dynamic model of a fixed-wing Unmanned Aerial Vehicle (UAV). A nonlinear observer is implemented to estimate vehicle states along with wind data such as angle of attack and sideslip angle. Simulation results show that the NAR- MAX approach has potential to estimate the correct model even when measurements is affected by noise, and in all cases produces a better model than the one provided by the Least Squares solution using a known model structure. Re- sults from the real world data set were found inconclusive but promising.
Important Links:
Go Back
