The Robots Are Coming: Autonomous Vehicles Could Be Next

There is something inevitable about robotics on the farm. Much of the work farmers and their employees do is repetitive and time-consuming, and does not require decision-making. Perfect stuff for robots, which don't get bored or tired, or demand a paycheck.
Kinze wirelessly links a tractor and combine. (Progressive Farmer image by Jim Patrico)
No surprise, then, that robots already have found a home on the farm. A prime example is DeLaval's VMS milking system, which hit the market way back in 1998. It uses an optical camera and lasers to "see" a cow's teats, a hydraulic arm to apply milkers and automated sprayers to clean the cow and send her on her way. If it didn't cost $250,000, it might have replaced all human milkers by now.
Other current helper robots include Case IH's new Robo-Sharpener, which rides on a rail to automatically sharpen discs on a tillage tool. How about drone airplanes that can scout fields with aerial photography? And the University of Illinois a few years ago produced a robotic weeder that could roll down 30-inch rows zapping weeds. (See Robots We Already Know.)
Such helpers are useful. But the Holy Grail of on-farm robotics might be autonomous vehicles. Theoretically, they could make a huge difference to a farmer's bottom line by saving him time in the field and money on labor costs.
Last year, Kinze Manufacturing took a stab at an autonomous vehicle when it unveiled a driverless tractor for pulling grain carts. Kinze gave dealers and the media a surprise glimpse of the technology at a meeting last July. Guests were ushered to a lawn at the company's headquarters as a combine came chugging from behind some buildings.
Alongside the combine was a tractor pulling a grain cart. The pair did a little maneuvering on the lawn as the combine unloaded on the fly into the grain cart. Then the tractor broke off and pulled the grain cart to a resting spot near the spectators. It wasn't until the tractor got near the spectators that the surprise became evident: The tractor had no driver. Cue the "wow" factor.
Kinze had worked for two years with Jaybridge Robotics, of Cambridge, Mass., on the project, which employed available technologies like GPS guidance and autosteer.
To use Kinze's system, a farmer will have to first GPS-map a field to determine where waterways, contours and other obstructions lay. Software can then plot efficient courses for the tractor and grain cart.
When the grain cart slides next to the combine during unloading, the driver of the combine must maneuver his vehicle to make sure the auger spout is over a partially empty area of the cart. The tractor can then independently take its load of grain to a staging area for unloading into a truck.
It is a slick system, and Kinze later announced a similar setup for a tractor/planter combo. The company said other autonomous applications were on the horizon.
What Kinze didn't announce was a release date for autonomous vehicle technology. "We need to figure out how we are going to commercialize it," Kinze COO Brian McKown said at the time.
Safety has to be one of the reasons for Kinze's delay in releasing the system. Although the system has sensors to make the tractor stop if it encounters obstacles, that might not be enough. For instance, can the sensors see to the side and predict that a moving object -- or person -- will intersect with the vehicle? One mishap with an autonomous vehicle could set the industry back for years.
"We have a lot of the right technology today. The only sensor we are lacking is the lawyer sensor," jokes John Posselius, Case IH innovation engineering director.
Liability and safety risks understandably make manufacturers nervous about unleashing a totally driverless vehicle. Still, Posselius says driverless farm vehicles will happen; the whole flow of technology almost dictates it. He recounts recent technological history that seems to lead inevitably to autonomous vehicles:
-- GPS guidance has made chores easier and less fatiguing. Over time, farmers have come to trust the technology.
-- Machines are learning to make automatic setting changes themselves, including concave settings and fan speeds in combines.
-- Tractor/implement interfaces are on the way, and it won't be long before the implement will control the tractor. For instance, a baler will know when the bale is ready to be wrapped, stop the tractor, wrap and eject the bale and start the tractor again.
Vehicle-to-vehicle (V2V) telematics have spawned a first generation of master/drone tractors. Case IH, Fendt and John Deere all have introduced systems that let the driver of one vehicle wirelessly operate another vehicle using GPS guidance. For instance, a combine driver can control a tractor pulling a grain cart. Or one tractor driver pulling a tillage tool can control a second tractor pulling another tillage tool. It would be an easy step to eliminate the driver of the second vehicle. But a second driver might add to the comfort level and might ease safety and liability concerns, at least until some technological bugs are worked out.
For instance, "RTK systems are super accurate. But there can be some problems with the base GPS systems," such as when there is a "hole" in the satellite constellations above a location or a signal skip or momentarily loss of signal because of atmospheric conditions, Posselius says.
And what if a steering system fails for some other reason? That's where redundancy backups take over.
"If there is going to be equipment moving without a person, it will have to be bulletproof," Posselius says.
In the end, he says, autonomous vehicles will find a place on the farm: "The only thing missing is time and a comfort level with the technology."
You don't have to look far to find robots or robotic vehicles. That fancy new system that lets your car parallel park itself? Robotics. That lawn mower your crazy neighbor has that bounces around the yard cutting grass? A robot. A decade ago, John Deere introduced a driverless tractor. While it has not yet come to market, it also is a robot.
Here are a few other robots that either are, or someday could be, on your farm:
Robotic blade sharpener. Case IH's Robo-Sharpener isn't high tech. But it could be a real labor saver. It runs on a trolley system from blade to blade on a tillage tool, tirelessly putting an edge on both regular and wavy concave blades on Case IH's True-Tandem 330 Turbo. It requires a human supervisor and an electric outlet for the motor to turn the blades with a gang-driver. (See it operate on YouTube at
Drone Airplanes Scout Fields. The military's drone airplanes are international newsmakers today. But civilians have been flying remote-controlled planes for decades. Now farmers are getting into the act.
Aerial images are a good way to scout fields for weed infestations, fertility problems and other issues you can't see from the ground. Last year, we reported on a couple of field-mapping services that employ small drone airplanes to take in-season aerial photos of fields. The cost for one service is $400 to $500 for photos of about 2,000 acres. True color images can spot a lot of problems from the air. But infrared and multispectral images have the potential to be even better tools.
Driverless Gator. It is not a farm vehicle, but John Deere does sell a robotic Gator. Eight years ago, it introduced the R-Gator ("R" for robotic) for use by the military. It goes where soldiers shouldn't ... as a scout and a perimeter lookout. It also can shuttle supplies from the rear to the front lines without a driver.
The R-Gator has front and rear cameras, a tilt and zoom camera that can see 360 degrees, and a night-vision thermal camera for work in the dark. An operator using a laptop, which displays up to four live video streams, controls it. GPS guidance allows the operator to know where the vehicle is at all times.
While the R-Gator is more at home in Afghanistan than Alabama, the technologies that make it possible are farm-ready. Who knows what the future holds for similar utility vehicles?
Robot Weed Killer. In the early 2000s, the University of Illinois made a name for itself as a robot breeding ground -- if that isn't a contradiction in terms. Agricultural Engineering Associate Professor Tony Grift led a series of graduate students into the study of robotics for the farm. They came up with trucks and harvesters and weed-killing machines.
The weed killers were especially advanced. They were created to roam the areas between rows of crops, where they would sense and then snuff out weeds with small squirts of glyphosate, mechanical saw blades or pincers.
"It worked about halfway," Grift says of the project. Then funding ran out, a graduate student graduated, and the project stalled out.
"The initial thrill [about robotics] is gone," Grift says. "But it will come back."


Publicar un comentario

Con la tecnología de Blogger.