Fully autonomous vehicles are closer than you think. You can now choose an autonomous vehicle from Uber and Lyft in select cities in the U.S., and while these vehicles come equipped with a driver for safety for now, the human safety net could be pulled as early as this year in some major cities.
Some are already doing it. Waymo, the Google-driven leader in this space, introduced a fully driverless ride-hailing service in Phoenix in late December. The cars have been available for a while to a small test group, but now anyone in Phoenix can ride in a Waymo vehicle without a human safety driver in the front seat.
Amazon-owned Zoox also unveiled its autonomous robotaxi in December. This fully driverless vehicle built for ride hailing doesn’t have a steering wheel or other human controls. It can go up to 75 mph and can run up to 16 hours on a single charge and is being tested in select markets.
More than half of U.S. states allow autonomous vehicles on the road, and national governments are addressing the issue in earnest, among them Singapore, Netherlands, China, South Korea, and Germany as leaders in creating national dialogue toward final standards for level-4 autonomy, which is full driving mode in limited area or geofencing. Level 5 autonomy, which does not require any human attention, is still contingent on policy and infrastructure.
MORE BY DAVID FRABOTTA
So What About the Farm?
The exciting part of this accelerated deployment — driven in part by COVID-19 and the need to mitigate human contact — is that a commercialized and deployed product will create best practices and innovation that can be shared by other cities and acclimate users to the new experience. In many ways, the technology is already a reality, and we’re just waiting for infrastructure, public policy, and consumer acceptance to catch up with the innovation that is already out there.
The same story is playing out on farms around the world as operators weigh the benefits and ROI versus personal preferences and regulation of autonomous features. For large-scale combines, harvesters, planters, spreaders, and sprayers from the major manufacturers, the technology for full autonomy might already be on-board and waiting to be used.
GPS-guided autosteer was the starting point and an example of how adoption began slowly and evolved into a ubiquitous feature that now comes standard on almost everything sold. Autosteer sets the baseline for technologies that are rolling out today.
“In 2016 we showed the moonshot of the cab-less tractor, and a lot of the components in that vehicle are making its way into our products today,” says Case IH’s Leo Bose, Advanced Farming Systems Marketing Manager for North America and Harvest Marketing Manager.
In the progression from GPS-guided steering to yield monitors, and variable-rate fertilizer application and seeding to capitalize on data about soil type, the constant through the adoption of each technology was ROI and integration into current cultural practices and acceptance that a technology was better than a current practice. It might take a year or two for farmers to accept an innovation, or it might take 10 years, but it is important to give users flexibility to integrate new technology slowly.
“You can see the progression today in what we call AFS AccuTurn, and it allows us to link into the guidance system so farmers can script out an entire field so when farmers start in a field, the operator will not need to touch that steering wheel, beginning, middle and end,” Bose says. “Some customers are still at the point where they want to control the turn on the end. But if you look at the algorithms that look at efficiencies, AccuTurn makes the most efficient and best agronomic choice for the implement you’re towing behind the tractor for that turn. So we still need to get better outcome information out there so farmers are comfortable relinquishing that control. If you look at AFS Harvest Command, which automatically adjusts through the field based on sensor feedback to increase or decrease rotor speed, sieve openings, or planning fan speed, customers are now more comfortable with the machines making those adjustments because it’s saving them money by putting more grain in the tank. That’s where we see the adoption coming about.”
Case launched Harvest Command in 2018, and today, Bose says about 90% of the machines it sells are equipped with the technology. That is a relatively fast adoption curve, which he attributes to the company’s customer-centric development. They know what functions are likely to be adopted quickly because farmers are already asking for help in specific areas.
Similarly, AGCO looks to innovate certain tasks with technology solutions based on farmer need, much of it centered around seeding.
“If you look at equipment farmers are using today, there’s a lot of smart capability already on those machines that you will see transferred or utilized in more autonomous or robotic platforms in the future,” says AGCO’s Matt Rushing, Vice President, Product Line, Global Crop Care.
AGCO owns Fendt, which developed the fully autonomous Xaver unit for precision seeding and weeding. While the concept platform has real-world potential, especially in swarms on high-value crops, the biggest benefit for many of these autonomous robots is what companies are learning from them so technology can be applied to machines that are more widely used.
“Smart solutions are being tested and tried today on some robot platforms by startups and even some larger players in the industry. There are emerging technologies and capabilities that you’ll actually see first deployed on a self-propelled machine that you could see on a typical farm today,” Rushing says. “And in a lot of ways, the autonomous platforms are giving us an opportunity to learn more about how smart solutions and sensors work, how machine learning works, what are the capabilities of some of these control systems, and then, immediately leverage them on machinery that farmers use today.”
There are three separate platforms that need further development to make level-5 machines that can diagnose, navigate, and act in the field:
- A decision platform interprets what you see, what to do about it, and how to react. It incorporates historical data, geographical data, weather, pest pressures, seed, soil, and other data that the farmer, consultant, and service provider use to make agronomic decisions.
- A carrier platform must then physically navigate to a place in the field to take action, which involves sensors, guidance, telemetry, obstacle avoidance, and many of the other technologies that we associate with autonomy.
- The application platform is then leveraged to execute a task, such as weed removal, planting, or spraying.
The decision platform has taken the longest to develop because it requires machines to learn best practices based on real-world recognition scenarios, but all platforms are accelerating more quickly amid heightened pressure on farmers, pressures that include access to labor amid immigration uncertainties, COVID-19 protocols, and increased production costs for inputs, minimum-wage increases, and regulatory and consumer pressure for more sustainable production.
“Many different factors are driving the faster adoption,” Rushing says. “Climate volatility, environmental and farm sustainability, regulations and restrictions regarding chemical use, all these things are driving faster adoption. If large chemical applicators cannot be used and we must use more site-specific weeding techniques, we must control those weeds mechanically. So having an autonomous solution that can run 24 hours a day, seven days a week consistently weeding a field is something that could be very attractive as an alternative to having costly manual labor perform the weeding task. These are examples of things that are driving farmers to investigate and become more accepting of automated systems faster.”
One of the areas that is progressing quickly is variable-rate crop protection. Throughout the development of automation guided by field imaging and analysis, fertilizer application and seeding have been areas that farmers adopted the quickest, largely because there was instant ROI to spreading less fertilizer and providing the best seed for the right environment. Crop protection has taken longer to develop and adopt, but the demand for more sustainable production methods and rising input costs are driving rapid change in the segment.
“If we look across the spectrum of where robotics are being deployed today, I think one of the key areas is in crop protection,” Rushing says. “If we look at the pressure that crop protection is under in regards to chemical use, specifically in places like Europe, there is a lot of pressure from a labor perspective. In some locations you either can’t use chemicals or their use is restricted, so typically manual methods are used to remove weeds or pests. That becomes very cumbersome, labor intensive, or very expensive. Pain points are becoming more visible in crop protection, and customers are starting to ask ‘How do I use robotics as a way to manage and protect my crop without having to use chemicals and beyond that, potentially even lower my labor cost?’
“We’re seeing some robotic and autonomous solutions being deployed into the marketplace already,” he continues. “And it’s great to see because we are starting to see through early feedback, what’s possible, what’s not possible, where farmer ROI is realized and where maybe it still needs some work. This will continue to happen over the next two to three years, and we believe it’s going to accelerate. And again, we will continue to see the first applications will be focused on commodities like high-value crops – vegetables, orchards, and vineyards are good opportunities to start – and see residual benefits beyond just labor and chemical reduction.”
Just this March, John Deere commercialized its anticipated select-spray technology. The AI-powered recognition system has the potential to only use a pesticide on individual pests in field. There are several select-spray technologies in development, many of which claim that the technology will reduce pesticide volumes 80%-90%.
Deere’s initial rollout is shy of the coveted green-on-green recognition like the technology it bought from Blue River in 2017, but just as other technology introductions, the volume of data and outcomes from the field will be invaluable in developing the next generation of the product.
Deere’s See & Spray Select uses camera technology to detect color differentiation in the field. As the sprayer moves through the field, its cameras rapidly detect only green plants within fallow ground and triggers an application to those plants. According to Deere, See & Spray Select has a similar hit rate to broadcast spraying while applying 77% less herbicide on average.
“There are a lot customers in the small grain area that have fallow in their rotation, and they might go over it two or three times trying to keep the weeds from growing to preserve nutrients and soil moisture,” says Joel Basinger, Marketing Manager for See & Spray. “This was a product we could get out now for them and fully integrate it into our sprayers to help them with that issue.”
This technology also makes it possible for farmers to use more expensive and complex tank mixes more efficiently than what they can achieve through a broadcast application, notes Basinger, reducing their costs and improving their ability to control herbicide-resistant weeds at a lower cost.
Built on the John Deere ExactApply foundation, a 400 or 600 Series Sprayer factory equipped with See & Spray Select enables farmers to have both a spot-spray and a broadcast machine in one integrated OEM package, says Basinger. See & Spray Select can be equipped on new MY22 and newer John Deere 400 and 600 Series Sprayers, which include the 408R, 410R, 412R, 612R, and 616R.
With the next model year of planters, John Deere is also ready to provide a fully integrated liquid fertilizer system to allow for the application of high rates of fertilizer at planting with the fewest of stops to fill and tender possible. These redesigned planters are part of a factory-installed, integrated ExactRate solution from John Deere that the company says eliminates the need for add-on tanks, custom plumbing solutions, and multiple suppliers.
“John Deere is the first company to bring to market a factory-installed, integrated liquid fertilizer delivery system on a tractor and planter together,” says Kevin Juhl, Marketing Manager. “A new MY22 John Deere Planter with increased capacities, paired with the integrated ExactRate tanks on an 8RX Tractor, is an evolutionary planting solution that can increase capacity, minimize soil compaction, and precisely place seed and fertilizer.”
With all the major manufacturers, level-5 autonomy isn’t necessarily the ultimate goal, instead it’s the result of incremental technologies that are deployed for farmers’ specific needs. The culmination of them, however, is making its way onto machines at a more rapid rate as farmers look for every possible advantage to meet food production needs that harmonize the benefits to people, the planet, and profitable farming enterprise.