PRACTICAL APPLICATION OF ADVANCED CONTROL : An evaluation of control methods on a Quanser AERO

Abstract

In the past years, rotorcraft have received much attention in our society. These provide great accessibility and are utilized in many areas of application. The complex dynamics of rotorcraft make them di cult to control, and the applications demand stable, safe, and agile devices. This thesis investigates the use of advanced control methods and evaluates their e ectiveness using a 2DOF (pitch and yaw) ' ight control experiment-platform'; the Quanser AERO. The main goal of the project is development and evaluation of angular-position controllers for the AERO, using conventional and advanced control theory. This includes cascade P-PI, LQR state-feedback, MPC and MRAC. The controllers are tested in simulation using linear and nonlinear models, and implemented for control of the AERO. The performance of each controller is evaluated in a standardized test. This test is combined step response sequence for pitch and yaw axes. The performance of each controller is evaluated through a performance index, which is based on the squared tracking error between command signal and plant output. This project has shown that model based controllers like LQR and MPC works well when tuned to a speci c case (for both simulation and practical application), but the performance crumbles when the system is changed without changes to the controller parameters. The simulation of control of a linear system for the conventional controller and MRAC gave excellent results, which con rms that the controllers work well in theory. When it comes to practical application, the conventional controller continues to work relatively well. The MRAC, however, shows strange behaviour, which was not anticipated. The results give insight to the practical application of advanced controllers, and inspires further research into the subject.