Advanced

Visual navigation and control of a mobile robot

Rudbeck, Fredrik LU (2017) In Master's Theses in Mathematical Sciences FMA820 20162
Mathematics (Faculty of Engineering)
Abstract
As cameras become cheaper and better, using them as sensors have become increasingly more common. This is especially true in robotics and for autonomous robot navigation. This thesis suggests a method where inputs from a single camera are used for making a robot self driving. The robot is assumed to move along planar surfaces, common in indoor environments.

The method is based on matching image features to a prebuilt navigation map also made up of image features. The method is evaluated by making a successful implementation on a mobile robot.

Since a camera is the only input, it proves that the quality of the camera is of high importance. All cameras in motion suffer from distortion. It is referred to as motion blur and arises when... (More)
As cameras become cheaper and better, using them as sensors have become increasingly more common. This is especially true in robotics and for autonomous robot navigation. This thesis suggests a method where inputs from a single camera are used for making a robot self driving. The robot is assumed to move along planar surfaces, common in indoor environments.

The method is based on matching image features to a prebuilt navigation map also made up of image features. The method is evaluated by making a successful implementation on a mobile robot.

Since a camera is the only input, it proves that the quality of the camera is of high importance. All cameras in motion suffer from distortion. It is referred to as motion blur and arises when the scene changes during the recording of a single exposure. To try and exploit that cameras are cheap, it is investigated if a simple USB camera is sufficient for the method. It proves to be, but to keep the motion blur low, the maximum velocity of the robot is heavily limited. (Less)
Popular Abstract (Swedish)
Robotar kan programmeras för att självmant röra sig runt i världen. I takt med att kameror blir både billigare och bättre är det vanligt att robotar använder kameror för att se sin omvärld.
Please use this url to cite or link to this publication:
author
Rudbeck, Fredrik LU
supervisor
organization
course
FMA820 20162
year
type
H2 - Master's Degree (Two Years)
subject
keywords
Image analysis, robot control
publication/series
Master's Theses in Mathematical Sciences
report number
LUFTMA-3311-2017
ISSN
1404-6342
other publication id
2017:E7
language
English
id
8905290
date added to LUP
2017-06-20 15:14:37
date last changed
2018-10-11 16:20:23
@misc{8905290,
  abstract     = {As cameras become cheaper and better, using them as sensors have become increasingly more common. This is especially true in robotics and for autonomous robot navigation. This thesis suggests a method where inputs from a single camera are used for making a robot self driving. The robot is assumed to move along planar surfaces, common in indoor environments.

The method is based on matching image features to a prebuilt navigation map also made up of image features. The method is evaluated by making a successful implementation on a mobile robot.

Since a camera is the only input, it proves that the quality of the camera is of high importance. All cameras in motion suffer from distortion. It is referred to as motion blur and arises when the scene changes during the recording of a single exposure. To try and exploit that cameras are cheap, it is investigated if a simple USB camera is sufficient for the method. It proves to be, but to keep the motion blur low, the maximum velocity of the robot is heavily limited.},
  author       = {Rudbeck, Fredrik},
  issn         = {1404-6342},
  keyword      = {Image analysis,robot control},
  language     = {eng},
  note         = {Student Paper},
  series       = {Master's Theses in Mathematical Sciences},
  title        = {Visual navigation and control of a mobile robot},
  year         = {2017},
}