A couple of years ago I was walking through the John Deere Pavilion in Moline, Illinois, when I noticed, parked on its own in its own corner, something I hadn’t seen in a good long time. It was the first autonomous concept tractor the green brand had ever shown to the public. Streamlined, roughly 60 horsepower, it was a utility tractor with no operator station that debuted briefly about a decade ago.
After that, there was a nine-year gap during which the brand didn’t promote the driverless concept or unveil any other fully autonomous tractor concepts for public until Agritechnica 2019 where it showed a unique tracked, autonomous tractor. But still just as a concept machine.
In between those two, though, the brand introduced a host of specific features on standard production equipment that were spun off from of its R&D on autonomous machine development, such as Machine Sync, the system that lets a combine operator control a tractor pulling a grain cart during unloading. It has essentially been easing the industry down a path toward a possible driverless future.
“We’ve been on this (autonomy) journey for quite a while,” said Margaux Ascherl, Deere’s program delivery manager for autonomy at the brand’s Intelligent Solutions Group (ISG). “I would say you’ve seen some of the product launches along the way benefiting from that technology. It’s in our guidance solutions; it’s in our data solutions. We’ve learned a lot as technology has grown along the way to enable some of this technology to be marketable and built into a product that is sustainable and farmers can actually use long-term. As technology has evolved, so have we.”
Ascherl spoke to Country Guide via internet conference in January as part of Deere’s effort to make direct contact with media members during its third appearance at the Consumer Electronics Show (CES). The CES is normally held in Las Vegas and highlights the latest in all manner of digital technology. Increasingly, big equipment brands like Deere are seeing themselves as a good fit for the central theme of that event.
Deere shocked the technology world in 2019, when it first parked one of its newest combine models in a prominent stall at that event. Staff informed showgoers about the artificial intelligence built into the machine and promoted agriculture in general as an industry already deeply committed to implementing cutting-edge technology.
That includes a future where autonomy will be prominent in machine design.
Three years on, in 2021, Deere was joined by Caterpillar, which wanted to show off its autonomous off-road mine trucks and talk about that technology, already being successfully used in open-pit mining. General Motors, too, used the show to make major announcements about the electrification of its future fleet, and about its continuing work on driverless technology.
“I was at the first CES event and one of the best parts of that was being able to tell people how much tech there was in ag,” Ascherl says. “That (feeding the world with technology) message is so easy to tell. And once we start telling it, there’s a lot of interest.”
Getting those spinoff elements from autonomous machine R&D into current production machines has also helped Ascherl and her teams get a better understanding of how autonomous technology can make a difference on farms. And, importantly, where it fits best.
“We’ve launched several products along the way that have benefited from that R&D,” she continues. “If you look at the ExactEmerge planter that can precisely place seed going seven m.p.h. or the see-and-spray products that we’re working on now and are publicly available. Our relationship with Blue River (a technology startup Deere purchased a few years ago) for example, with machine learning. And our Combine Advisor. Those are good examples where we’ve found a nice way to introduce technology that makes sense on a customer’s farm now, because that’s what really matters. A lot of those technologies have benefited from the autonomy journey along the way and will continue to do so.
“(With) those early concepts, we learned a ton, but one of the biggest learnings from those projects was we’re not doing autonomy for the sake of autonomy. We’re doing autonomy because it’s one solution to customers’ problems.”
Now, says Ascherle, Deere is looking across its entire fleet to find where autonomous technology will have the best fit. And, she stresses, that means the best fit with farmers’ operations, not with Deere’s product line. If autonomous is going to succeed, she says, farmers have to see that it will work for them.
“Managing an autonomous fleet shouldn’t create more problems around labour, supervision and understanding how a machine is working and what it’s going to do — all of those key components. What I think we’ve learned along the way is how important that is.
“With new technology there’s unintended consequences. And sometimes there are unintended benefits. We learned how those technologies would impact their (farmers’) businesses and their farms overall, and take that into consideration. Even some of those products we’ve released so far that are not fully autonomous. We’re taking some of that customer feedback and integrating it into our autonomy strategy as well.”
As Deere — and others — ease farmers into a higher-tech future, Ascherl doesn’t think the lack of 5G connectivity in rural areas will be an impediment to development, although it likely would help farmers who want to take full advantage of what engineers can create for them as use of on-farm technology continues to grow.
“I don’t think it’s necessarily a holdup to (autonomous machine) development. It limits some of the technology you can introduce and some of the connectivity,” she says. “It doesn’t necessarily stop it (development).
“We’ve actually tackled it and started looking into localized solutions for real rural areas. (But) increased connectivity will always help. It will help farmers generally, across their entire fleet and therefore autonomy as well. It’s one of those technologies that enables more. So with autonomy it really depends on the task. But I wouldn’t say any place that doesn’t have full connectivity couldn’t do autonomy.
“I think we’re still learning about how necessary 5G is. But the faster you can push data, it’s always going to make things easier. But I wouldn’t rule out autonomy and say that it’s not capable until we have 5G. But we’re still looking into it and thinking: Without it, what can we do?”
During Deere’s official online presentation at CES, which was intended in large part for those who don’t know much about agriculture, a spokesman said he absolutely expects to see full autonomy in a number of machines come into widespread use on farms during his lifetime.
Ascherl sees a much shorter timeline.
“I think we’ll see it way sooner” she says with a broad smile. “It’s not too far off.”
In fact, it’s just over the horizon, she says. “Where are we in the journey? I can’t wait until we can talk about the projects that we have in hopper.”
Student engineers look to ag
Although true on-road driverless vehicles capable of travelling anywhere any time remain tantalizingly out of reach, autonomous machine operation in off-road sectors like mining and agriculture is already gaining market acceptance.
“(Off-road) you’re operating in an environment without lots of people,” says Margaux Ascherl, John Deere’s program delivery manager for autonomy at the Intelligent Solutions Group. “You don’t have to go at speeds of 70 m.p.h. Stopping is easy in agriculture, whereas it’s not on a highway. It doesn’t require the same level of infrastructure to invest in as self-driving cars will. Ag, itself, provides a great environment for autonomy to go first.”
No surprise then that university engineering students looking for a career in technology are seeing the benefits of turning their emerging skills to projects designed for farming. Despite not having a background in agriculture, a group of students at the University of British Columbia in Vancouver have decided to do exactly that.
Senior students in engineering fields there have embarked on their AgroBot project.
“UBC AgroBot is a student design team,” says Aviral Salhotra, image recognition lead. “So it’s completely owned by students, mostly from the engineering department. We’re trying to build a robot that can do targeted elimination of weeds in a farm field. We think targeting the weed and spraying pesticide only at the locations where you need it, rather than over the entire field, is a much more sustainable solution.”
“We actually don’t come from a farming background,” adds Devika Vishwanth, navigation team lead. “It was something one of our friends was interested in when we started this project. We thought it was a very good application of the technology we were learning about in school.”
With almost two years of R&D work behind them, the student team had its sights set on participating in Purdue University’s AGgrowBOT competition, originally scheduled for this coming May. However, it’s been cancelled due to the COVID-19 pandemic.
The UBC team is split into two main groups: the image recognition group looking to create software that can distinguish crop from weeds and the navigation group focused on guidance.
Planning for the Purdue event required the team to create a program that could work in fields of winter wheat, something not easy to find in the Greater Vancouver area.
“We’re trying to train on some other crop for now and later translate it for wheat when we get more data,” explains Salhotra. “We’re going out to the UBC farm. We’re talking to them to see what crops they’re growing there.”
The team’s robot uses two cameras, one for plant identification and one for guidance. By late fall the machine was still in the prototype stage and the students continued to work on it. And so far the team is working without any corporate sponsorship.
“We are contacting some companies now,” says Vishwanth. “We may be incorporating some more technologies soon, I think. But right now, we’ve been operating on our own.”