Development of a visual-teach-and-repeat based navigation technique on quadrotor aerial vehicle

Nguyen, Trung (2014) Development of a visual-teach-and-repeat based navigation technique on quadrotor aerial vehicle. Masters thesis, Memorial University of Newfoundland.

[img] [English] PDF - Accepted Version
Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission.

Download (8Mb)

Abstract

The objective of this thesis is to develop a vision-based navigation and control technique for quadrotor to operate in GPS-denied environments. The navigating technique has been developed while using Visual-Teach-and-Repeat (VT&R) method. This method is qualitative where the position of the quadrotor is estimated based on a set of reference images. These reference images are collected while taking the quadrotor manually along a desired route. Each image, collected in the database, represents one segment of the desired route. The features are extracted from these images using a well-known method, Speeded-Up Robust Features (SURF) [1]. When the quadrotor is navigated along the desired route (repeat mode), the quadrotor performs self-localization. Three methods of self-localization are presented. In method I, the SURF features observed on the current image are matched with the SURF features of the reference images to compute the probability value of each segment in the desired route. The segment that provides the best probability value is chosen as the current segment of the quadrotor. To improve the accuracy of localization, in the method II, the condition of feature-size relation with spatial distance is imposed. In the method III, the estimation of the current segment of the quadrotor is built on Bayes’s rule. Based on the appearance-based error of feature coordinates, the system computes qualitative motion control commands (desired yaw and height) for the next movement in order to control the quadrotor to follow the desired route. This computation is developed on Funnel Lane theory, which was originally proposed in [2], in order to 2D navigate ground vehicle following the desired route. The thesis extends it to 3D navigation for the quadrotor. Funnel Lane theory qualitatively defines possible positions where the vehicle can fly straight by the constraints of features coordinates between the current image and the reference image. If the quadrotor locates outside the funnel lane, it will be navigated back to the funnel lane. A nonlinear geometric controller has been developed to convert the motion control commands, generated basing on VT&R technique, into control inputs necessary for the four rotors in the quadrotor. The design of proposed controller is simplified by concentrating on the errors of rotational matrix, instead of attempting to access the errors of each degree of freedom. The quadrotor for this thesis is chosen as the well-known AR.Drone model [3]. The whole system is modeled and simulated in Gazebo simulator using Robot Operating System (ROS). Four image databases have been used for testing self-localization: two databases around Engineering building of Memorial University of Newfoundland, COLD database and New College database. With proposed VT&R technique, the quadrotor is able to independently follow a long route without GPS-information or the support from an external tracking system. The proposed system has a simple implementation, inexpensive computation and high potential for exploring and searching-and-rescuing missions.

Item Type: Thesis (Masters)
URI: http://research.library.mun.ca/id/eprint/8283
Item ID: 8283
Additional Information: Includes bibliographical references (pages 95-105).
Department(s): Engineering and Applied Science, Faculty of
Date: October 2014
Date Type: Submission
Library of Congress Subject Heading: Quadrotor helicopters--Automatic control; Navigation (Aeronautics); Robot vision; Image processing; Computational intelligence

Actions (login required)

View Item View Item

Downloads

Downloads per month over the past year

View more statistics