Control and navigation of a quadrotor subject to wind disturbance

Abstract

Recently, the use of small-scale rotorcraft unmanned aerial vehicles (UAVs) for surveillance and monitoring tasks is becoming attractive. Their usage can be extended for monitoring of oil and gas pipelines. Amongst the various configurations of small-scale rotorcraft UAVs, the use of a quadrotor gained more prominence, particularly in the academic research community. A quadrotor is a small responsive four-rotor vehicle controlled by the rotational speed of its rotors. It is compact in design with the ability to carry a high payload. Achieving a successful outdoor surveillance/monitoring task with a quadrotor requires a robust and disturbance-rejection flight controller. However, to design a robust disturbancerejection flight control system for a quadrotor, the understanding of its aerodynamic permanence is essential. The fact that thrust is the key variable which influences the overall performance of rotary-wing air-vehicles, whilst wind velocity variation affects thrust utilization, their overall impact on a quadrotor manifests a rise or a fall in the rotors thrust which results in a non-uniform rotors’ thrust, and subsequent drop in altitude. This thesis addresses the effect of wind velocity variation in the design of a quadrotor autonomous flight control system. The following represents the main achievements recorded in this study: extensive review of the relevant literature related to UAVs with emphasis on rotorcraft, a quadrotor system was designed and developed in-house, followed by its aerodynamic analysis using the computational fluid dynamics (CFD) modeling technique, where ANSYS FLUENT and a virtual blade model (VBM) were employed to analyse the wind velocity variation effects on the quadrotor. A frequency-domain system identification technique was used to extract the quadrotor parameterized model using the comprehensive identification from frequency responses (CIFER) software package. Matlab/Simulink tools and Arduino-Simulink blockset were utilized in the design of the quadrotor’s altitude and attitude regulating controllers, based on the classical PID and model predictive control (MPC) schemes. Two algorithms were also developed, one for the vehicle’s navigation control and the other a graphical user interface (GUI) which facilitates semi-autonomous/autonomous control of the quadrotor. A series of actual flight tests were conducted to evaluate the performance and effectiveness of the quadrotor system. The tests involved careful selection of specific missions, based on which corresponding flight trajectories for the flight scheduling layer in the flight control structure were defined and the actual flight tests conducted. The flight scheduling involves ii the generation of appropriate flight trajectories for certain defined manoeuvres to evaluate the robustness and handling quality of the overall flight control system within the vehicle’s operational envelope. The defined manoeuvres include ascend-hover-descend, forward flight, hover turn and circular-zigzag flight. Each flight manoeuvre was defined in terms of clear objectives, full description and performance requirement, which was categorized into two qualitative levels, namely desired level (satisfactory) and adequate level (barely acceptable). Prior to conducting the extensive flight tests, the responses of the vehicle to specific directional control commands was examined using a frequency sweep excitation signal. The results of these tests were considered more than acceptable. This study lays a foundation for related researches, particularly in the development of small-scale rotorcraft flight control systems from an aerodynamic point of view. KEY WORD: ANSYS FLUENT, Arduino-Simulink blockset, CIFER, Flight control system, Flight control software, Monitoring, MPC, PID, Quadrotor, Surveillance, Wind velocity variation.v