| Abstract
|
:
|
Unmanned Aerial Vehicle (UAV) flight control systems must be capable of delivering the required performance while handling nonlinearities and uncertainties in the vehicle model, the atmosphere, and ambient wind. These factors necessitate the development of nonlinear flight control system design methods that can handle large nonlinearities and uncertainties. Variable approaches to the linear control of UAVs have been discussed in the recent literature. However, the development of a nonlinear robust autopilot has not been addressed to any significant degree. The development of a nonlinear autopilot based on robust control methods will be discussed in this dissertation. In this design technique, the nonlinear UAV model is not linearized. The control law is designed using the H∞ technique. This dissertation presents the results of an exploratory study to examine robust autopilot nonlinear design methods for the UAV and compare this new approach with existing PID, LQR, and linear H∞ techniques. Since the method must then be verified, its flight simulation will be done using MATLAB/SIMULINK. Verification, validation and robustness tests are documented at the end of this dissertation. The airplane examined is called the Hawkeye. It was designed and built by KU students in the fall of 2004. It is a small, 14 foot wingspan, remotely controlled airplane made from composite materials with a maximum takeoff weight of 90 lbs. It will be used in the future as a small UAV for research programs at KU. The mission includes take-off, climb, cruise, a one and a half circle accomplished in a level turn, and a return back to its original airfield accomplished by cruising back, descending, and completing an approach and landing. After take-off, the airplane is required to climb to 1,000 ft altitude, and then it travels 5,000 ft over the ground into the target area. It will then take some photos of that target using its camera. The complete mission for the UAV lasts about 180 seconds.
|