Hover and forward flight of an Autonomous UAV using optimal linear control and gain scheduling

Bæk, Teis ; Hammelsvang, Mads ; Jørgensen, Thomas Bæk

10. term

Elektronik og IT, Kandidatuddannelsen (Spec. Intelligent Autonomous Systems)

2007

2007-06-07

0 pages

This master's thesis concerns the design of linear control strategies for making an AUV capable of autonomous flight. The goal is to be able to complete level 1 in the IARC, which involves flight over a distance of 3 km, as fast as possible. The starting point for this thesis is taken in the developed nonlinear model of a reconfigured Bergen Industrial Twin helicopter, where the linear controllers are designed based on linearization of the nonlinear model. Two different linear control design methods are used for the controller design; classic SISO control and optimal LQR control. Furthermore, a gain scheduling approach to perform scheduling between two controllers is introduced. As a high level controller a supervisory controller is designed to handle position control. Hover control was designed using the two different linear control design methods. The classic SISO controllers was implemented as lead or lag compensators. Furthermore, a forward flight LQR controller was developed to be able to perform fast forward flight. The gain scheduling approach was implemented as an observer-based gain scheduler framework around the LQR hover and forward flight controllers. The LQR hover and forward flight controllers, along with the observer-based gain scheduling controller, were all tested separately in three different level 1 specifications to be able to compare them. It was concluded that the observer-based gain scheduling approach was the most time efficient control method of the three.

Download
Download
Download
View record in AAU Student Projects