NCSU Institutional Repository >
NC State Theses and Dissertations >
Please use this identifier to cite or link to this item:
|Title: ||Agricultural Robotics Using Absolute Position Sensors on a Zero Turning Radius Platform|
|Authors: ||Powell, Nathaniel B|
|Advisors: ||Gary Roberson, Committee Member|
Mike Boyette, Committee Chair
Larry Stikeleather, Committee Member
Doug Barlage, Committee Member
|Issue Date: ||4-May-2007|
|Discipline: ||Biological and Agricultural Engineering|
|Abstract: ||Autonomous field equipment which can successfully replace human operators on tractors and other equipment may revolutionize agriculture. Replacing current field equipment such as large tractors with autonomous machines designed so that one person can supervise multiple machines will make it possible to use more machines without increasing the number of human operators needed. This makes it feasible to use smaller machines, which have benefits for cultivation, safety and the cost of purchasing and maintaining the equipment.
This research is a step in the process of developing autonomous equipment. Using a zero turning radius vehicle, a prototype autonomous vehicle was constructed which measures position with a differential global positioning system (DGPS) receiver. A real time kinematic GPS (RTK-GPS) receiver was used for development and testing to obtain the most accurate results available. This work built upon previous autonomous navigation work using different sensors including a laser positioning system and a machine vision system.
In addition to the GPS receiver an electronic compass was used to aid with turning behavior. These sensors were coordinated using inexpensive microcontrollers as input processors and a third microcontroller to coordinate the inputs and control the machine by controlling electronic actuators installed on the steering mechanisms. A simple control algorithm was implemented by combining the crosstrack and angular deviations from the reference path and applying a control input to the actuators based on that combined or "aggregate" error. The control input was calculated using a the traditional Proportional-Integral-Derivative controller.
The system was developed and tested on a flat grassy lawn bordered by trees and buildings which provided an ideal simulation of a landscape management application. A straight line path 45 meters in length was set up for testing the machine. The machine started at one end of this path and navigated to the other end. The performance was measured by recording the position reported by the RTK-GPS on a handheld computer and calculating the crosstrack error between the recorded position and the straight line path.
The results demonstrated that the autonomous navigation system was able to navigate the machine from point to point. The performance did not meet the desired standards of precision and was characterized by oscillations which were of higher amplitude and higher frequency than is acceptable for agricultural machines. Limitations in the microcontrollers' ability to store and manipulate the position data were a major factor. However, the potential of small zero turning radius platforms for agricultural robotics is significant and the simplicity of controlling such vehicles is a great advantage over traditional tractors.|
|Appears in Collections:||Theses|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.