Note: This content is accessible to all versions of every browser. However, this browser does not seem to support current Web standards, preventing the display of our site's design details.


Vision Based MPC on a Low-Cost Quad-Copter


Christian Saner

Master Thesis, FS13 (10289)

This work aims for autonomous exploration in an unknown indoor environment with a quad-copter. In this scenario, no external sensors such as GPS, motion capture system or arti cial landmarks are available. Therefore, position and orientation estimation must be provided by on-board processing of proprioceptive and exteroceptive sensors. The system developed in this thesis is based on an AR.Drone 2.0 frame. The on-board electronics are replaced by the PX4 system provided by PIXHAWK. An Odroid U2 board provides the computational power for image processing, sensor fusion and model predictive control. A newly developed visual odometry algorithm that is able to process 40 frames per second is used. Since the state estimation is based on visual information it is crucial that the control is not too aggressive, ensuring a stable camera image at all times. Therefore an MPC with state constraints is introduced to maintain limits in the camera movement and therefore guarantee the required image quality. The experiments conducted reveal that preprogrammed ight gures such as squares and eight-shapes can be tracked robustly. The precision is comparable to results in related work. The developed system constitutes a good basis for further research in autonomous exploration. iv


Type of Publication:

(12)Diploma/Master Thesis

S. M. Huck

No Files for download available.
% Autogenerated BibTeX entry
@PhdThesis { Xxx:2013:IFA_4542
Permanent link