(Dr. Heichel)
(Robert Dunker)
& Thanks
University of Illinois at Urbana-Champaign logo

Automated Guidance for Row-Crop Cultivation

John Reid John Reid
Department of Agricultural Engineering
(217) 333-2738; j-reid1@illinois.edu
Qin Zhang
  Qin Zhang
Assistant Professor
Department of Agricultural Engineering
(217) 333-9419; qinzhang@illinois.edu

The potential for vehicle automation in North American agriculture has increased by the advent of precision agriculture. The modern agricultural tractor is capable of controlled guidance by utilizing electrohydraulic systems and controller area networks on the vehicle to provide easy access to steering control. The development of guidance technologies potentially provides the benefits of high accuracy, higher-speed operations, reduced fatigue, and increased operator safety. The University of Illinois has recently completed a C-FAR program to evaluate guidance technologies for Illinois crop production. This presentation reviews the progress in such technologies.

Numerous posture-sensing systems were evaluated, including mechanical feelers, GPS, RTK-DGPS, machine vision, geomagnetic direction sensors, and inertial sensing systems. Additionally, combinations of these sensing systems have been evaluated. Mechanical feelers, machine vision, and machine vision combined with GPS are valid sensors for vehicle guidance.

Mechanical feeler guidance systems are commercially available and utilized by a number of Illinois farmers. These technologies are effective for straight row guidance performed at moderate speeds (less than 6mph). These systems had limited success when used on curved rows, at higher speeds, or any time when the plants are not sufficiently strong for the mechanical interface with the sensor. A typical system costs about $2,500.00.

Machine vision technology uses a vehicle-mounted camera and computer processor to detect crop rows. Machine vision is an effective sensor for straight or curved rows and has been proven successful at speeds of 11 mph. A machine vision guidance system performs much like the "cruise control" on automobiles. The operator engages the system with a single button and the system tracks crop rows across the field. The operator takes control of the system when the brake is depressed, the steering wheel is turned, or when the end of the row is detected. While no commercial systems are currently available, the cost of the basic hardware is competitive with mechanical feeler guidance systems for tractors having electrohydraulic steering.

Some sensors used individually have limited performance and sometimes fail. Sensor fusion provides a software method to incorporate the signals of multiple sensors to provide a more accurate guidance signal. We have combined the signals of machine vision, RTK-GPS, and inertial sensors to provide a robust system that has been tested at speeds of 7 mph and has accuracy of better than 2 inches. The system primarily uses machine vision sensing when the crop stand is good, and GPS guidance from a map under adverse crop conditions.

Detailed information on the performance of the various sensors is available upon request.

The Wired Tractor

Back to Agronomy Day 2000 Field Presentations Index

Department of Crop Sciences
College of Agricultural, Consumer and Environmental Sciences
University of Illinois Extension
Copyright © 2000 University of Illinois
Email site problems to the webmaster
Site Map