Development and validation of a controlled virtual environment for guidance, navigation and control of quadrotor UAV

Abstract

This thesis is focused on the development of a six degrees of freedom (6DOF) simulation model of a commercial-off-the-shelf quadrotor. The dynamics of the quadrotor and its control strategy are described. The Geometric Dilution of Precision (GDOP) of the Autonomous Systems Engineering and Integration Laboratory (ASEIL) laboratory used in conducting the experiments is also analyzed. Simulation results are then verified with actual flight data. A direct method of calculus of variations is employed in the development of an algorithm for optimal trajectory generation and collision-free flight. Using the differential-flatness characteristics of the system, the trajectory optimization is posed as a nonlinear constrained optimization problem in virtual domain, not explicitly related to the time domain. Appropriate parameterized functions employing an abstract argument, known as the virtual arc, are used to ensure initial and terminal constraints satisfaction. A speed factor maps the virtual to the time domain and controls the speed profile along any predetermined trajectory. An inner loop attitude controller was used to achieve almost global asymptotic attitude tracking for trajectory following. The trajectory generation and following algorithms were verified using the 6DOF simulation model through a simulated collision avoidance scenario.