Ag Autonomy Grows Ever Bigger With John Deere CES Announcements

These days, driverless cars are transporting millions of people in China and the U.S. Those riders probably don't realize that the food they eat may have benefitted from autonomous vehicles as well!

Three years after unveiling an autonomous tractor at CES 2022, the John Deere Company revealed new products at this year’s show that apply autonomy to an ever-growing number of tasks needed to plant, protect, maintain, and harvest a crop. The company’s second-generation autonomy kit combines advanced computer vision, AI, cameras, and lidars to help the machines navigate their environments.

While upgrading and expanding their Ag offerings, Deere is expanding autonomy to landscaping and heavy hauling at resource extraction sites such as quarries. While each of these industries experience their own set of challenges, the common factor across all is lack of skilled labor. The American Farm Bureau Federation estimates that there are roughly 2.4 million farm jobs that need to be filled annually. In construction, eighty-eight percent of contractors struggle to find skilled labor. And in commercial landscaping, 86% of landscaping business owners can’t find labor to fill open positions.

“Our agriculture, construction, and commercial landscaping customers all have work that must get done at certain times of the day and year, yet there is not enough available and skilled labor to do the work,” said Jahmy Hindman, Chief Technology Officer at John Deere. “Autonomy can help address this challenge. That’s why we’re extending our technology stack to enable more machines to operate safely and autonomously in unique and complex environments. This will not only benefit our customers, but all of us who rely on them to provide the food, fuel, fiber, infrastructure, and landscaping care that we depend on every day.”

Fields

Tillage is one of the busiest times of the year for large-scale farming. For Deere’s Autonomous 9RX Tractor the second-generation autonomy kit features 16 individual cameras arranged in pods to enable a 360-degree view of the field. Farmers can step away from the machine and be more productive focusing on other important jobs. The advanced autonomy kit also calculates depth more accurately at larger distances, allowing the tractor to pull more equipment and drive faster.

Two Nvidia processing chips, running at 50 watts steady state, handle the image processing computing from 16 cameras, notes Igino Cafiero, Director of Bear Flag Robotics within John Deere. “At a top speed of 12 mph, this is a 40% speed increase for our new autonomous tractor,” he said.

Orchards

Orchards are the latest frontier in Ag Autonomy, in particular almonds and pistachio farming. Here, unique challenges exist in keeping the crop free of pests. This is accomplished via an air-blast pesticide spraying process, a challenging and repetitive job which has to be done at least six times during a growing season. Spraying occurs at night while the pests roost in the branches. The drivers must wear hazard suits while moving at only 2 mph. All in all, a miserable job.

Deere’s Autonomous 5ML Orchard Tractor (shown above) incorporates their latest autonomy kit. Because GPS signals can be blocked by the dense leaf canopies found in orchards, map-only based operations can be unreliable. Deere’s solution adds a lidar sensor for localization and obstacle avoidance.

Quarries

Quarries supply the essential raw materials vital for building roads, buildings, and infrastructure. It’s a complex process to mine, process, and transport materials. The 460 P-Tier Autonomous Articulated Dump Truck (ADT) handles the repetitive tasks of transporting material around the quarry to facilitate different steps in the cycle.

The ADT is 34 feet long and 12 feet high and can carry over 92,000 lbs. of material in its bucket.

The mining autonomy space has been active for years; Caterpillar and Komatsu were early to the game and startup SafeAI is now testing their retrofit offering in a U.S. quarry. But there’s a wide variety of machines used in specific settings and operational approaches. With their ADT, which can operate in a variety of operations in addition to quarries, Deere seeks to create a niche for themselves.

Grounds

Commercial landscapers are tasked with the upkeep of office parks and industrial campuses. These operations are highly labor intensive. It is a competitive space where work is typically awarded on a competitive basis. Once a contract is signed, having the staff to support new work is essential. Mowing is a low skill task compared to tending to flower beds and tree health. With limited staff, managers see autonomy as a means of applying their staff resources more productively.

Deere’s autonomous commercial electric mower leverages the same camera technology as other Deere autonomous machines, but on a reduced scale since the machine has a smaller footprint. With two cameras each on the front, left, right, and rear, 360-degree coverage is achieved, allowing staff to focus on other aspects of the overall job.

It can operate up to 10 hours per cycle, which aligns well with a typical shift. Another advantage of electric propulsion is that noise pollution is reduced and jobs can start earlier in the morning.

Deere revealed that the data generated from deployment of their first-generation tractors enabled AI-based development of their construction and mowing products.

Autonomy Support

Deere says that select machines will be autonomy ready from the factory and the second-generation perception system will be available as a retrofit kit for certain existing machines, providing customers with multiple paths to adoption.

The machines are managed via John Deere Operations Center Mobile, the company’s cloud-based platform. Through the Deere app, users can start the machine and also have access to live video, images, data and metrics, and the ability to adjust various factors like speed.

And, similar to the remote assist processes used by leading on-road autonomy players, In the event of any job quality anomalies or machine health issues, users will be notified remotely so they can make necessary adjustments.

A Sharp Autonomy Dividing Line

Deere’s new autonomy offerings are not the only play in agriculture. Competitor AGCO has introduced a variety of field-oriented autonomous machines. Vertical farming is another highly automated tech path.

For agriculture, Deere and other similar offerors have been focused on supporting customers when “making the product,” i.e. a crop. This aligns with indoor automation which is used extensively in the manufacturing of goods.

For any producer, vital supply chain factors kick in once the crop is harvested or the gadget plops off an assembly line. For manufacturers, this is when trucking or other modes of transportation kick in for transporting the product to the point of sale. Farmers need to transport a harvested crop from the field to a grain elevator or processing plant, typically on public roads. Many farmers doing the shipping are Deere customers: what an opportunity to expand into a new domain!

However, at a tech demo for journalists preceding CES, Deere made it clear their product boundary is the edge of the field or orchard. For understandable reasons, of course. In the fields or in the orchards, all equipment is part of the operation; all movements are directed by the grower. On road, all equipment (vehicles) operate independent of one another. Laws and norms govern their actions to a degree but the human element is unpredictable. Crashes can happen. The legal and insurance factors create real headaches and must be built into the pricing transport services.

In what I call the “Ag Adjacent” supply chain, the same labor shortage exists and autonomy offerings are popping up to serve this need.

For example, in North Dakota, the Minn-Dak Farmers Cooperative grows and processes sugar beets. MDFC is one of America’s largest sugar beet shareholder/grower cooperatives. The sugar beet harvest period is very compressed, about three weeks. Finding thousands of drivers to stop whatever else they were doing to come fill this need is problematic.

In 2022, MDFC began working with Kratos Defense to include the use of autonomous vehicles in its operations. “Leader/Follower” truck platooning is deployed to haul harvested sugar beets between piling stations and granulated sugar processing plants in Minnesota and North Dakota. A driver in a lead truck is followed by an autonomous truck, such that a single driver is moving two full loads. For now, both North Dakota and Minnesota regulations require those backup trucks to have a "safety driver" behind the wheel who can intervene to operate the vehicle if something goes wrong. Based on several years of operations, the proponents of this approach are seeking to allow an unmanned follower truck.

A similar operation is in advanced stages in northern Canada, where the crop is trees. Moving timber from the cutting area to a sawmill occurs on remote unpaved roads. It is a challenging job and not enough drivers are available to meet the need. Forest Product Innovations, a research group funded by the Canadian forestry industry, has been working with Forterra to implement leader-follower platooning so that one driver can move more freight.

Labor Unions?

By staying “inside the fence,” Deere has it easy compared to the on-road autonomy players who must deal with powerful labor unions. The Teamsters in particular are highly active in state legislatures, pushing non-sensical bills to require that truck drivers be on-board autonomous trucks. So far, no such legislation has passed, but slowing down the deployment of autonomous trucks is clearly a priority for them.

But labor unions exist for workers in landscaping, farming, and quarry operations as well. Quarry workers can be members of the United Steel Workers of America. Farm hands can belong to the United Farm Workers of America, and The Laborers' International Union of North America represents construction craft laborers, including those in landscaping.

It’s worth watching to see how labor interests will play into these job markets. Since these operations are seemingly always short of workers, the opportunity is for workers to be able to transition to more skilled (and higher paid) work roles than simple driving.

What’s Next?

Deere’s goal is to have a fully autonomous farming system for corn and soybeans in the U.S. by 2030.

At the same time, Deere is taking a measured approach to deploying autonomy. Only several dozen of their autonomous tractors announced in 2022 are in daily operations with farmers, with extensive interaction with Deere technical experts to evaluate and tweak performance. A Deere spokesperson at CES said that this “hand-holding” process allows the product to be thoroughly evaluated and hardened prior to full availability. Makes sense. I assume this type of slow rollout will be the case for this year’s new products as well. Compared to autonomy startups who need to move quickly to product to generate revenue, this is a viable and prudent option open to a well-established market leader such as Deere.

How will these new Deere products evolve?

For tree crops, what are the remaining manual jobs that could be addressed by autonomy, such as harvesting the crop?

For the autonomous landscaping product, what adjacent uses cases exist? Possibly the maintenance of large parking lots, for sweeping and snow plowing?

In quarries and other types of resource extraction, the very competitive market across CAT, Komatsu, Hyundai, Volvo Autonomous Solutions, and now Deere will continue to innovate, as the market pull is strong.

What’s clear is that autonomy will keep seeping into every nook and cranny of the outdoor work space.

Disclaimer: Richard Bishop advises and/or holds equity in the follow companies mentioned in this article: Forterra.