- 352 views
- 724 downloads
UAV Linear Model Predictive Control Using Computer Vision Algorithms
-
- Author / Creator
- Surma,Christopher
-
This thesis aims to develop a motion control strategy for an Unmanned Aerial Vehicle (UAV) to execute a pursuit algorithm based on a vision based object detection algorithm. This enables a pursuer UAV to follow a target UAV based on images obtained from the onboard camera of the pursuer. The UAV pursuit algorithm is implemented onto a commercially available Parrot AR.Drone 2.0 quadcopter. Two motion control strategies considered for the UAV pursuit algorithm are a Propor- tional Integral Derivative (PID) Control and a Linear Model Predictive Control (LMPC). The performance of these two control strategies is evaluated based on their performance responding to a step input in each input channel, and tracking a figure-8 flight trajectory. A LMPC strategy was chosen to follow the target drone’s trajectory estimated by a vision based object detection algorithm. In order to use a LMPC strategy, a linear state space model was obtained and identified for the Parrot AR.Drone 2.0. The vision based object detection algorithm used for our application is YOLO v2, a single-layer convolutional network which identifies the location of the target drone and returns a bounding box around it in the image frame. Experimental testing proves the proposed UAV pursuit algorithm achieves accurate detection of the target and successfully pursues it using the LMPC motion controller.
-
- Graduation date
- Fall 2019
-
- Type of Item
- Thesis
-
- Degree
- Master of Science
-
- License
- Permission is hereby granted to the University of Alberta Libraries to reproduce single copies of this thesis and to lend or sell such copies for private, scholarly or scientific research purposes only. Where the thesis is converted to, or otherwise made available in digital form, the University of Alberta will advise potential users of the thesis of these terms. The author reserves all other publication and other rights in association with the copyright in the thesis and, except as herein before provided, neither the thesis nor any substantial portion thereof may be printed or otherwise reproduced in any material form whatsoever without the author's prior written permission.