Machine Design and Vision Based Navigation
Gautam, Samrat (2014)
Gautam, Samrat
Hämeen ammattikorkeakoulu
2014
All rights reserved
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2014072313342
https://urn.fi/URN:NBN:fi:amk-2014072313342
Tiivistelmä
This study covers the design of an autonomous robot and its testing
process on an artificial maize field constructed for an indoor environment.
However, the ultimate goal of this project was to participate in the Field
Robot Event 2014 organized by the University of Hohenheim in Germany.
This project was commissioned by HAMK University of Applied
Sciences. And was fabricated and tested in the automation laboratory of
HAMK UAS valkeakoski.
The test result obtained by plotting the signal from wheel encoder, sonar
sensor, gyroscope, magnetic compass was used as a primary source of information.
Different scientific literature published on four wheel differential
drive and vision based navigation was well examined for background
information. Beside literature, rules and the regulation of the FRE 2014
were used as a source of information as well. In addition to these a working
video on a previous field robot event provided a good reference for
planning and designing an autonomous robot.
A four-wheel differential drive chassis with a suspension system was designed
and fabricated. Sensors such as a magnetic compass, gyroscope,
sonar sensor, wheel encoder and camera were used to sense the environment.
A suitable control algorithm was developed to meet the requirements
of the competition. An indoor test field was designed with the artificial
maize plants made up of paper and plastic tube. This test field was
used to examine the control modules designed for different level. After a
series of testing and tuning, a smooth navigation through row of corn was
achieved. Oscillation was dropped down to nominal level; obstacle was
identified from a safe distance and weed plants was successfully detected.
The findings suggest that the performance of the robot is satisfactory.
However, there are several possibilities for improving it. These include:
replacing a sonar sensor with laser range scanner for detecting maize
plants. Simultaneous localization and mapping can also be introduced. The
author strongly recommends implementing a laser range scanner and stereo
vision in to a future project.
process on an artificial maize field constructed for an indoor environment.
However, the ultimate goal of this project was to participate in the Field
Robot Event 2014 organized by the University of Hohenheim in Germany.
This project was commissioned by HAMK University of Applied
Sciences. And was fabricated and tested in the automation laboratory of
HAMK UAS valkeakoski.
The test result obtained by plotting the signal from wheel encoder, sonar
sensor, gyroscope, magnetic compass was used as a primary source of information.
Different scientific literature published on four wheel differential
drive and vision based navigation was well examined for background
information. Beside literature, rules and the regulation of the FRE 2014
were used as a source of information as well. In addition to these a working
video on a previous field robot event provided a good reference for
planning and designing an autonomous robot.
A four-wheel differential drive chassis with a suspension system was designed
and fabricated. Sensors such as a magnetic compass, gyroscope,
sonar sensor, wheel encoder and camera were used to sense the environment.
A suitable control algorithm was developed to meet the requirements
of the competition. An indoor test field was designed with the artificial
maize plants made up of paper and plastic tube. This test field was
used to examine the control modules designed for different level. After a
series of testing and tuning, a smooth navigation through row of corn was
achieved. Oscillation was dropped down to nominal level; obstacle was
identified from a safe distance and weed plants was successfully detected.
The findings suggest that the performance of the robot is satisfactory.
However, there are several possibilities for improving it. These include:
replacing a sonar sensor with laser range scanner for detecting maize
plants. Simultaneous localization and mapping can also be introduced. The
author strongly recommends implementing a laser range scanner and stereo
vision in to a future project.