In recent months, John Deere has aggressively promoted its latest technology in ads that feature an experimental driverless tractor. Deere says its tractor will help engineers “bring space technology back to Earth.” By pairing agricultural navigation tools with experimental space exploration technologies developed at Utah State University, the company has developed what may be the most advanced (and downright weirdest-looking) driverless tractor to date.
Deere's experimental tractor has no cab, steering wheel, seat or anything else to accommodate an on-board driver. It's designed specifically as a spray tractor for orchards, pointing to a couple of reasons why a farmer might someday want to buy a driverless tractor.
Merv Kizlyk, Deere's manager for worldwide advanced research, notes that a driverless sprayer tractor could reduce a farmer's exposure to insecticides. And since it doesn't need daylight, it can just as easily spray at night, when wind conditions are typically more favorable to spraying. Meanwhile, the trees and low-hanging branches in an orchard provide an obstacle course for the experimental tractor to navigate.
The green machine uses multiple vehicle control and navigation systems to adapt to changing and harsh farming environments. Camera-computer configurations, radio receivers and infrared sensors let the machine see though dust, darkness and even thick crop canopies. A wireless control module allows the farmer to control the vehicle from a distance and change its settings, reroute it, start it or shut it down.
“By overcoming real-life farm navigation challenges, we learn how to better use existing technologies and see where new technology needs to be developed,” Kizlyk says. “What we learn from this experimental tractor may end up as new technology in many types of new tractors in the future. I wouldn't say we're producing a robotic vehicle, but rather a smart vehicle that can react to change, go around obstacles and return to its original path. With that in mind, it's easy to see how farm applications can be every bit as challenging as situations faced by an automated soil sampling vehicle on Mars. Human tractor drivers perform many functions that are difficult for a smart machine to duplicate.”
Think of the possibilities
John Deere Senior Engineer Dan Niebher says one practical future autonomous application for Midwestern row-crop farmers might be an automated grain cart. “As the combine grain tank gets full, it could automatically send a signal for a driverless grain cart to come out and pick up the load,” Niebher says. “Another application might be a master-slave setup where one tractor in the field has a driver who remotely controls several driverless smart vehicles in the same field. This could work for tillage, harvest, spraying or any operation that might require multiple farm vehicles.”
The potential benefits of such vehicles are limited only by the imagination. Eliminating the need for a cab saves weight and cost, labor shortages become less of a concern, operator fatigue is reduced or eliminated and human safety may be improved. Still, one of the biggest challenges for Deere and other manufacturers is to determine how much technology farmers will be willing to pay for.
“Before investing millions of dollars in research, we conduct farmer focus groups to determine if anyone might be interested in buying something like this,” Niebher says. “The answer we consistently get is that farmers will buy cutting-edge technologies, but only if they can see a definite economic payback. So for every nickle's worth of technology, we've got to produce a dime's worth of value.”
Kizlyk points out, too, that not every farmer is interested in high-tech tractors. “A significant number of farmers simply want a reliable tractor with lots of power; they don't care about technology. Some don't even like digital readouts or anything electronic associated with their machines. We have to be mindful not to alienate those customers as we're chasing down better technology for the early adopters who always want something new.”
Back from the future
Exactly when we'll see driverless tractors crawling across fields in great numbers is still a matter of debate, with university engineers and industry experts predicting as few as three years to as many as 10 years or more. However, most of the engineers agree that technology revolutions such as driverless tractors happen when enabling technologies converge.
In the world of tractor automation, technologies converge in many ways. Universities publish papers, large companies work with small companies, and as with Case and New Holland, large companies merge. According to Abe Orbach, senior director of strategic technology development for Case/New Holland (CNH), often the greatest challenge is seeing how to piece together different technology components into something useful. In the process, the engineers might also produce something flashy.
Robotic tractor hope and hype had already started to build well before last year's CNH merger. In 1999, New Holland and Carnegie Mellon University began sporting around a driverless self-propelled windrower. The machine used GPS guidance and VisionGuidance (a technology that pairs cameras and computers) to “see” differences in the crop line. The machine harvested 100 acres of alfalfa without stopping.
Despite the success of experimental prototypes such as the autonomous New Holland windrower, Orbach says he's more interested in the immediate, practical applications of new technology. “We look at driverless tractors more as laboratories on wheels than as something farmers will want to buy anytime soon. Our main focus is on practical uses of technology. We'll continue to build on our Advanced Farming System, for example. It's been available on combines since 1996 and was recently added to planters. This is a tremendous tool for site-specific agriculture applications such as linking to GPS to keep the machine driving in the same wheel track each time, accurate plant spacing and variable rate seeding. The system is easy to install and integrate into the complete cropping system from planting to harvest.”
Looking to the future, Orbach says that equipment technology from both Case and New Holland will be designed to ensure that every implement can communicate on the same network. He notes that Magnum and STX tractors already come wired to accept Control Area Network systems that improve communication and data transfer between computers, satellites and machines. “I think we'll soon see GPS-enabled systems used to control many more automatic functions on machines,” Orbach says. “Possibilities range from variable rate seeding to tractor hitches that automatically raise the implement at the end of the row.”
Drive by wire
Dexter Schaible, senior vice president of engineering and development for AGCO, also sees today's driverless machines as more of a research and development tool than something that farmers might buy anytime soon. “Taking the driver out of an experimental tractor forces engineers to think of the tractor in different ways,” Schaible says. “For example, you quickly realize how much weight and cost might be saved by eliminating the steering rods and mechanical steering linkages of a conventional tractor. One way to accomplish that task would be to use a drive-by-wire system where the on-board human driver steers via radio signals instead of a mechanical steering system. Aircraft pilots have used fly-by-wire systems successfully for years, and automobile manufacturers are looking at replacing mechanical steering in cars with radio transmitters and servos, so there's no reason we couldn't do something similar with wheeled tractors. Someday, we might even replace the steering wheel with a joystick.”
In fact, drive-by-wire tractors are already a reality. John Deere's T-series “tracked” tractors use drive by wire to control the vehicle's steering system.