Development of an intelligent model prediction controller for autonomous helicopters

Kubali, Şevket Eser
In this thesis, a new PID gain update law using linear least squares regression is introduced as an adaptive control method for autonomous helicopters. In addition, future prediction analyses are conducted for error dynamics of the closed loop system using recursive linear least squares regression. Combining these two concepts with classical PID controller, an intelligent PID controller is obtained. On the other hand, using PID controllers, a flight controller with three control loops is developed to demonstrate the capabilities of the new intelligent controller and PID controllers of second and third control loops of this flight controller are replaced by the newly developed intelligent controller. Thus, a new intelligent flight controller is acquired with model prediction and adaptation abilities. Several challenging maneuvers are carried out in virtual environment for the flight controller that has no adaptation ability and the new intelligent flight controller using the same initially stable PID gains to investigate the success of the new intelligent controller.
Citation Formats
Ş. E. Kubali, “Development of an intelligent model prediction controller for autonomous helicopters,” M.S. - Master of Science, Middle East Technical University, 2016.