Precision Ag Aids Sustainability
Deere VP Deanna Kovar, a 2021 SAE COMVEC keynoter, provides thoughts on autonomy's future, computer vision, machine learning and data management.
Ten years ago, some “visionaries” at John Deere made a critical decision: that every one of its agricultural vehicles needed to be an IOT device that connected to the cloud. That way, agronomic and machine data could flow more easily to help farmers make better decisions in real time.
“When we made that decision, it wasn’t easy to literally put additional costs in every machine we built. But we had some visionaries who understood that in the future, this data would help us to help farmers farm even better,” Deanna Kovar, VP of Production & Precision Ag Production Systems, Deere & Co., shared with SAE Media. “Over time, we have continued to build opportunities for customers to get value from that data.”
Kovar will kick off the SAE COMVEC event , taking place September 14-16 in Chicago and online, by presenting the opening keynote address. She will discuss how highly automated and IoT-connected vehicles enable farmers to improve productivity and sustainability. Precision ag and the work Kovar’s group is pursuing ties well into the 2021 event’s theme – Energy, Sustainability and the Future of Commercial Vehicle Transportation.
Kovar recently spoke with SAE Media to preview her COMVEC presentation, providing insights into a range of connectivity-related topics.
Can you talk about the latest technology in precision ag?
Obviously, we’re continuing to stay on the trend of 3G, 4G and 5G technology changes. It’s exciting that 5G will drive more opportunities for farmers and farms to be connected. In rural America, we struggle with connectivity. We’re huge proponents of not thinking about how many people there are to connect in rural America, but how many acres there are to connect because connecting every acre has tremendous value.
The other part of this is really moving from being reactive to proactive because we can get this data literally in near real time. When we started this process 10 years ago, we were sending customers alerts that something failed; now we’re using AI [artificial intelligence] to instead say something’s about to fail. With prognostication and AI, we’re able to plan for that event because in farming, timing is so important. If you miss your planting window by a day, it’s 1% in yield. Or if you wait too long to harvest your soybeans, you can lose 1% of yield as well. So every day matters, every hour matters.
How does precision ag tie into sustainability?
One of the most exciting things for me about agriculture is that the customer’s profitability is not counter to sustainability practices. There’s a lot of energy going into how we can help farmers be more productive and grow more food and fiber with fewer inputs. That’s the definition of sustainability and it’s also how farmers can be more economically sustainable in their operations as well. So, we’re spending a lot of time thinking about sustainability on the farm and how technologies within precision ag and beyond can help that.
What do you hope attendees take away from your COMVEC keynote?
Part of it is helping the attendees understand how we’re leveraging technology in agriculture. I think it’s a longstanding misnomer that farmers aren’t sophisticated individuals. At the heart is just sharing, what is precision agriculture? Why is it so important for everyone on the planet? And how we’re thinking about the next generation of technology. We’ve been at this precision ag business for multiple decades, and we’ve made a lot of progress. There are items around data privacy, user experience, automation and autonomy where agriculture is in a unique seat to be able to provide insights to other industries about how we’re implementing technology at scale.
So other industries can look to the ag industry as a leader in these areas?
Relative to data privacy, for example, we collect very sensitive information from farmers. Every fall we get their yield data, which is basically their topline revenue information, on an acre-by-acre level. We have been very transparent with the industry about the farmer being in control of that data and very transparent about what we will and what we won’t do with the farmer’s data, putting them in control of who gets to see it, how John Deere gets to use it. We just want to help them collect and visualize and get insight from their data and connect them to over 150 connected software companies that are helping to close the full loop for farmers because farmers don’t farm alone. It’s not just Deere and the farmer, it’s their agronomists, their landlords, their banker, their input providers. All of these people must come together to help them drive these sustainability goals, both economic and environmental.
Where do you see precision ag heading in the next decade?
Precision ag over the past two decades, the 2000s were really all about GPS and auto guidance – self-guided machines through the field. The 2010s were all about data management. With cloud and connectivity, it really took a ton of information that was locked in desktop computers and put on binders – literally printed maps on binders and shelves in farmers’ offices – and moved it to the cloud. That created a whole new opportunity for easy data flow, easy visualization and insight creation for farmers. The 2020s are really all about computer vision and machine learning. How can we continue to augment the farmer’s senses with additional sensors and cameras, and then use the power of machine learning and artificial intelligence to sense and ultimately act on behalf of the farmer?
Deere bought a Silicon Valley-based company called Blue River Technology in 2017. They’re focused on exactly that, on leveraging computer vision and machine learning to differentiate between a weed and a healthy plant. We’re on a path to reduce herbicide applications by as much as 80 or 90%, because instead of broadcasting we’re literally managing at the plant level and only spraying the weeds. If you think about agriculture, we’ve gone from managing at the farm level to the field level, to the subfield level, because of GPS, automation, because of cloud-based computing and data management. The next decade is all about moving from the subfield level to plant level management, making sure that every seed gets placed perfectly, making sure that every weed is killed and every plant is nurtured individually. And then at the end of the season, we can harvest the entire crop efficiently.
Do you see much technology sharing with the on-road commercial vehicle sector for autonomous solutions?
Part of the reason that agriculture will be able to drive autonomy into the future is because vehicles will help us keep the costs of the infrastructure required in check. Agriculture also is in a unique position that maybe allows us to do it at scale faster because we are managing in very distinct plots of land without a lot of people running around. From a safety perspective, we know we have to borrow perception systems and obstacle-avoidance systems from other industries, but that we possibly have opportunities to use them at higher scale.
Now, the thing that makes it harder for us is we’re not just trying to get from point A to point B safely. We’re trying to do that and execute a job – placing hundreds of seeds per second at two inches deep, two inches apart from each other, which isn’t an easy thing to do. There’s a lot of instinct and monitoring that the operator does in the cab of a tractor or combine or sprayer today. Agriculture can’t move to autonomous solutions until we’re confident that they can do the job as good if not better than an operated vehicle. That challenge is a big one for us, because there’s a tremendous amount of automation of jobs that we’re executing beyond just driving.
What’s the timeline for when autonomous systems approach the same level as human operations?
Some jobs are easier to do it on than others. In the coming years we will see some agricultural jobs become autonomous-ready. There are solutions out there already. In 2001, we demonstrated an autonomous orchard tractor, so we’ve been ideating about this for two decades. Yet, it’s not here. But because all of the technology converging – persistent 5G connectivity, the cloud, GPS – and because of all the automation we’ve already been doing that’s valuable even with an operator in the seat, we’re getting closer to it every day.