Content area
Abstract
This thesis presents measurement, estimation, and control schemes to aid a quadrotor unmanned aerial vehicle (UAV) in landing on a flat, inclined surface without prior knowledge of the surface's inclination. The system uses a single CMOS camera and several inexpensive laser modules for onboard sensing to measure the distance to and orientation of a landing surface. A nonlinear least squares estimation scheme yields the altitude of the quadrotor and the normal vector defining the ground plane. This information is used to design a hybrid landing trajectory composed of a position tracking phase and an attitude tracking phase. A geometric nonlinear control system is used during each phase and ensures that the quadrotor's attitude is aligned to the inclination of the ground surface at touchdown. A quadrotor is developed from the ground up to test the in-flight measurement process and to execute landing trajectories on an inclined surface. Experimental results demonstrate the quadrotor's ability to accurately estimate altitude and ground plane orientation during flight, and numerical simulations of landing trajectories for various surface inclinations are validated by experimental results up to a maximum inclination of thirty degrees.
