Over the next ten years, can John Deere join Tesla and other Silicon Valley tech behemoths as one of the top AI and robotics firms in the world?
This idea would seem out of place given the common misconception of the 185-year-old business as a heavy-metal maker of tractors, bulldozers, and lawnmowers with the company’s distinctive green and yellow paint jobs.
According to Jorge Heraud, vice president of automation and autonomy for Deere, based in Moline, Illinois, that is what the company envisions for the future. A preview of this was shown at the Consumer Electronics Show in Las Vegas in January of last year when Deere unveiled its fully autonomous 8R farm tractor, which is operated by artificial intelligence rather than a farmer at the wheel.
Deere’s nearly two decades of strategic planning and investment in automation, data analytics, GPS navigation, internet-of-things connection, and software engineering have culminated in the autonomous 8R. While a significant portion of that R&D was conducted in-house, the business also made a number of acquisitions and formed partnerships with agtech companies, gaining access to their know-how and expertise in the process.
This is the result of Heraud’s and his team’s insight that technology will boost agricultural output, profitability, and sustainability.
We are very, very, very early in this process, according to Stephen Volkmann, an equities research analyst at Jefferies, despite the fact that Deere made a major impression at CES and captured the interest of the investing community.
Less than 50 Deere tractors are currently autonomously driven globally, he continued. Furthermore, Volkmann said that while Deere intends to have a completely automated farming system for row crops in place by 2030, “in Wall Street time, it’s an eternity.”
Deere is now making money and adding value with well-proven automated technologies that can be adapted to its current tractors, such GPS-based self-steering and precision seeding that calculates how deeply and widely to space plants. Before you can give kids full autonomy, Volkmann said those conditions must be met.
Without even mentioning the commercial advantage, the autonomous 8R marks a significant advancement in contemporary agtech. “Everyone believed [complete autonomy] was pie in the sky until its debut at CES,” said Scott Shearer, chair of the Ohio State University department of food, agricultural, and biological engineering.
Although none of the autonomous tractor ideas are now commercially accessible, Shearer estimated that there are about 30 distinct ones being worked on throughout the world. But when Deere, which holds 60% of the North American tractor market, introduces one, “reality kicks in,” Shearer said.
This situation is a reflection of Deere’s autonomous plan. “The AI we deploy combines computer vision and machine learning,” Heraud added. These fields of study were actively being pursued at Silicon Valley company Blue River Technology, which Deere acquired in 2017 for $305 million. With the help of dozens of advanced cameras and processors, Blue River’s “see and spray” robotics platform can discriminate between agricultural plants and weeds when dispensing herbicides.
Six stereo camera pairs that are mounted on the autonomous tractor can “see” an impediment in the field, such as a rock, a log, or a human, and estimate its size and distance. A deep neural network assesses each pixel in an image taken by the cameras and determines whether the tractor should go forward or stop in around 100 milliseconds.
In order for the tractor to utilize machine learning, Heraud explained, “we’ve selected hundreds of thousands of photographs from different agricultural areas and under varied weather and lighting circumstances. The farmer can also use this feature to remotely control the tractor while doing something else, rather than having to sit on it.
Heraud was alluding to autonomous driving, another element of Deere’s agtech jigsaw that was completed when the company paid $250 million to acquire Bear Flag Robotics last year. The automated navigation technology from Silicon Valley firm Bear Flag may be put onto pre-existing tractors. The most recent version of the Deere 8R tractor, which first hit the market in 2020, incorporates Blue River technology to have autonomous capabilities.
Deere has bought AI assets from two other leading agtech companies since the CES release. In April, Deere and GUSS Automation, a company that has created semi-autonomous orchard and vineyard sprayers, established a joint venture. A single operator may remotely manage up to eight GUSS (Global Unmanned Spray System) sprayers at once while using a laptop thanks to AI and IoT. Regardless of height or canopy size, GUSS can locate trees and calculate the appropriate amount of spray to apply to each one.
A month later, according to The Robot Report, Deere announced the purchase of multiple patents and other intellectual property from AI company Light. By employing extra cameras and simulating the anatomy of the human eye, Light’s depth-perception platform enhances existing stereo-vision systems and enables more precise 3D vision. The technology from Light will be included into next iterations of Deere’s autonomous agricultural machinery.
Deere has set up a Startup Collaborator program to test cutting-edge products with customers and dealers without a more formal commercial engagement in order to keep a close watch on other agtech R&D. In order to retain them in the fold, Volkmann added, “the aim is that they locate the jewels before they become evident to [competitors].” Four Growers, a Pittsburgh-based startup that offers robotic harvesting and analytics for high-value crops, beginning with greenhouse tomatoes, and Burro, a Philadelphia-based company that makes small, autonomous robots that can help farm workers with various conveyance tasks, are among the current crop.
Unsurprisingly, Deere’s main rivals have been working on agricultural machinery automation and autonomy as well. According to Seth Crawford, senior vice president and general manager of the Duluth, Georgia-based company’s precision agricultural and digital business, AGCO, whose brands include Massey Ferguson and Fendt, “has been automating farming operations since the mid-1990s.” We are currently in a phase of machine autonomy that we refer to as supervised autonomy, he explained. “Fully automated operations are the talk, but farmers are prepared to pay for automation on a feature-by-feature basis,”
Crawford stated that although Deere is concentrated on giving its own agricultural equipment complete autonomy, AGCO is eyeing a larger retrofit market. We’ll have a performance-improving retrofit kit accessible for many brands of machines in the summer of 2023, he promised. He made reference to Deere’s 8R and added, “Where others claim we provide you autonomy with a half-million-dollar tractor, we have kits that allow you to achieve it with your existing fleet. Farmers who want to use technology to improve their results but don’t want to switch their entire fleet and make that significant investment represent a major market potential, in our opinion.
A vehicle known as the Autonomous Concept Vehicle was brought to the Farm Progress Show in 2016 by Case IH, a division of London-based CNH Industrial. The modern, driverless prototype tractor gave away the idea of autonomy at the time. Six years later, Case IH introduced its Trident 5550 autonomous applicator at the Farm Progress Show in September.
The 2017 release Trident 5550 is a cab-equipped spreader for dry and wet materials in agricultural areas. The agricultural show model was updated with Raven Industries’ autonomous technology, which CNH purchased in June 2021 for $2.1 billion. The improved Trident uses self-driving technology, cutting-edge cameras, and AI to evaluate a constant stream of photos in order to detect impediments, much as Deere’s autonomous 8R.
According to Chris Dempsey, worldwide director at Case IH Precision Technology, the business intends to make a small number of the equipment available for farmers to test before going on sale sometime next year. The precise release date is still to be established. Before we launch, he stated, “We want to get user feedback and understand their trust level [in autonomy].”