Robust Control of Vectored Thrust Aerial Vehicles via Variable Structure Control Methods

摘要

The popularity of Unmanned Aerial Vehicles (UAVs) has grown rapidly in many civil and military applications in the last few decades. Recent UAV applications include crop monitoring, terrain mapping and aerial photography, where one or several image sensors attached to the UAV provide important terrain information. A thrust vectoring aerial vehicle, a vehicle with the ability to change the direction of thrust generated while keeping the UAV body at a zero roll and pitch orientation, can serve well in such applications by allowing the sensors to capture stable image data without additional gimbals, reducing the payload and cost while increasing the flight endurance. Furthermore, thrust vectoring UAVs can perform fast forward flight as well as hover operations with non-zero pitch: features which can serve well in military applications. The first part of this research focuses on developing a comprehensive dynamic model and a low level attitude and position control structure for a tri-rotor UAV with thrust vectoring capability, namely the Vectored Thrust Aerial Vehicle.Nonlinear dynamics of UAVs require robust control methods to realize stable flight. Special attention needs to be given to wind gust disturbances, and parametric uncertainties. Sliding Mode Control , a type of Variable Structure Controller, has served well over the years in controlling UAVs and other dynamic systems. However, conventional Sliding Mode Control results in a high frequency switching behavior of the control signal. Furthermore, Sliding Mode Control does not focus on fast set-point regulation or tracking, which can be advantageous for UAVs and many other robotic systems.Taking these research gaps into account, this work presents an Adaptive Variable Structure Control method, which can acquire fast set-point regulation while maintaining robustness against external disturbances and uncertainties. The adaptive algorithm developed in this work is fundamentally different from current Adaptive Sliding Mode Control and other Variable Structure methods. Simulation and experimental results are provided to demonstrate the superiority of the proposed approach compared to Sliding Mode Control. The novel adaptive algorithm is applicable to many nonlinear dynamic systems including UAVs, robot arm manipulators and space robots.The same adaptive concept is then utilized to develop an Adaptive Second Order Sliding Mode Controller. Compared to existing Second Order Sliding Mode Control methods, the proposed methodology is able to produce reduced sliding manifold reach times and consume less amount of control resources: features which are particularly advantageous for systems with limited control resources. Simulations are conducted to evaluate the performance of the proposed Adaptive Second Order Sliding Mode Control algorithm.