Your Reading List

Doing the robot math

The fully autonomous Greenbot made its commercial debut at Agritechnica in November.

In my last post I talked about the scale of some of today’s largest farm machines and what the future may—or may not—have in store for them. As technology advances, however, the question is becoming how much longer will we see very large scale equipment at work on broad-acre farms.

The reason I say that is because autonomous machine development has been steadily progressing. Basically, I’m talking about robots. This past week I had a chance to talk to a few engineers who are working on advancing the development of ag robots.

Premiere farm machinery shows, Agritechnica in particular, have seen autonomous machines appear with increasing frequency, and at least one company was displaying a market-ready machine at that show last November.

It’s no longer a matter of wondering if robots are going to become commonplace on prairie farms, it’s when and what will they look like. The chances are they will be small compared to today’s high-horsepower machines. Instead of one large unit operating on a farm, the thinking in some robotics circles today is to have a “swarm” of small machines working together.

The conversations I had this week with those engineers working on autonomous systems centred around their entries in the upcoming AgBot Challenge that will take place in Illinois in May. The Challenge offers a U.S.$50,000 cash prize for first place.

What it takes to grab that cash is to develop a machine that will load itself with seed and plant a dozen rows of corn in the field, all without the help of a human. That’s far from an insurmountable task according to the engineers.

I had a chance to see one of the robots, which is being developed by a team of engineering students at the University of Regina. It’s based on an off-the-shelf 26 horsepower tractor and will plant two rows at a time. The automated control system being designed to run it can easily be scaled up to work on a tractor of any size. And it, too, uses off-the-shelf components.

But will it make sense to scale it up to operate very large machines, or would a fleet of relatively small machines that can run as long as there is fuel in the tank and seed in the hopper be a better strategy? That is the swarm idea, using a fleet of smaller machines to get the job done. 26 horsepower, though, may be too small—or maybe not. But how about 50 or 100?

Remember, a driverless tractor doesn’t need to stop for coffee breaks or sleep at night. Lets do a bit of math to see how acreage numbers could add up if non-stop swarms were at work.

If a single small machine seeded a width of only 10 feet at 5 m.p.h., it could cover about 6 acres/hour. Say 4 to allow time for turns and fills, etc. That’s 96 acres per unit in a non-stop day. Multiply that by just 4 machines and you’ve seeded 384 acres in just one day. For the median sized family farm on the prairie, that’s a pretty respectable day’s work.

But consider what happens if those little machines could seed at 10 m.p.h. That’s not far fetched; there are some planters on the market now that can do that. The hourly tally goes to just over 12 acres/machine. Say 8 to be really conservative. With that level of technology the daily acreage tally goes north of 768.

At that rate, seeding a 3,500 acre farm would take only about 4 ½ days. Naturally, that doesn’t take into account time spent moving between fields, doing maintenance or any other reason to stop work.

Now go online and price out 4 small open-station, front-wheel assist tractors with hydrostatic transmissions (to make retrofitting robotic control systems easier) and compare that to one very-high horsepower, four-wheel drive tractor with a comfortable cab that is only going to work 12 hours a day, maybe 14. And remember, there’s no need to pay for air ride seats and leather wrapped steering wheels on the drones.

The technology already exists to do this. One Manitoba farmer, Matt Reimer, has built his own autonomous control system from scratch and put a driverless tractor to work on his farm. He’s not even a trained computer programmer or engineer. He’s willing to help others do the same with his new company, Reimer Robotics.

The cost-benefit analysis for each farm is going to be more complex than this simple exercise, but it’s clear from this look, there can be benefits.

What’s really mind blowing is it’s not beyond the realm of possibility to start to move to some level of autonomous field operations now.


About the author


Scott Garvey

Scott Garvey is a freelance writer and video producer. He is also the former machinery editor at Grainews.



Stories from our other publications