John Deere Bets the Farm on AI, IoT
Over the past 100 years, John Deere, with its iconic green tractors, has been a serious technology investor.
It started in 1918, when Deere bought the Waterloo Gasoline Engine Company, transforming the company from a farming implement business into a fully fledged industrial tractor firm. Fast forward to the late 1990s, and Deere invested in GPS with its acquisition of Navcom, which helped pave the way to installing 4G LTE modems on all its equipment.
Now, two decades into the 21st century, it's artificial intelligence, machine learning, neural networks, 5G and the Internet of Things (IoT). In 2017, Deere paid over $300 million for Blue River Technology, a Silicon Valley startup that is applying these technologies to the agricultural business. (See John Deere Is a Machine Learning Company Now.)
As the senior vice president for Deere's Intelligent Solutions Group, John Stone not only oversees the Blue River acquisition, but the company's other technology innovations, including a greater reliance on infrastructure-as-a-service (IaaS) to support its cloud computing needs.
"You can just think of ISG as the high-tech division of John Deere, so we have the responsibility to develop, design and embed these new technologies into all of our equipment," Stone told Enterprise Cloud News during the recent Mobile World Congress expo in Barcelona, where he was interested in the latest developments in 5G and IoT.
"When you say IoT, you normally think of things that fit in your pocket," said Stone, who has worked for the company for 16 years and has a background in mechanical engineering, as well as programming. "The 'T' for us is ten-ton tractors. Our large equipment now has 4G LTE modems, with WiFi and Bluetooth, and that does two-way communication so it collects data off your farm and sends it to the cloud. It also takes instructions from Deere or from dealers or other software companies and sends that to the machine. The two-way communication tells the machine what to do. Also, machines can also communicate with one another in the field."
The Blue River deal is expected to take this approach to industrial machines and farming to the next level.
Now that the deal between Deere and Blue River is closed, the two companies are starting to collaborate more, although Stone notes that Blue River will continue to operate as an independent firm.Future of farming
Blue River is developing a machine learning-based technology called "See and Spray," which can reduce the amount of herbicide used in farming. Specifically, a sprayer takes pictures of plants and by using machine learning and algorithms it can determine which ones are weeds and which ones are crops, and only spray herbicide on the weeds.
That's important when a typical field can have between 1 and 2 million plants.
This approach to farming offers a two-fold solution. The first is reducing or eliminating the amount of herbicide sprayed on the food people eat. The second is cost reduction. Stone estimates that a soybean or cotton farmer could save $50 to $80 per acre, and reduce herbicide spending up to 90%.
This is only beginning.
"Our roadmap is calling for machine learning and AI to find their way into every piece of John Deere equipment over time," Stone said. "What we do with our eyes can be done more accurately with a camera and a computer, with a system that retains that data and never forgets, and gets smarter with every pass of the field. This also applies to our construction and heavy equipment divisions, too."
In addition to AI, machine learning and IoT, there are investments in automation. The most obvious example is self-driving machines. However, there are other advances as well. Later this year, Deere plans to release an app that will offer machine setting adjustments from a central location, and the operator can either accept or decline those suggestions.
As more and more of this is automated, a central location can operate multiple machines in the field, making suggestions and relaying additional information.
Underpinning all this is the cloud, specifically Amazon Web Services Inc. . In addition to using Amazon's IaaS platform, Deere is a heavy user of Lambda, AWS's version of serverless computing that allows cloud apps to respond to different events, thus using resources only when needed. (See Serverless Computing: Why You Should Wait.)Deere's tech future
The next big step for Stone and his team would be the widespread introduction of 5G to speed up more of the processes, and create greater bandwidth to support more sophisticated applications and collect greater volumes of data -- high-definition video, for example -- whether it's from the machines or the field itself.
For now, with fully fledged 5G rollouts still some time off, Deere contents itself with the 4G LTE modems attached to every vehicle, which allows edge computing and the offloading of data to the cloud. (See A 5G Smartphone Timeline for 2019.)
As in the past, Deere is not limiting these technological developments to itself, although it remains the primary user. Instead, these developers are likely to be marketed and sold to other companies, making Deere a tech vendor in its own right.
"We know for sure that tech is the future, and that AI and machine learning is the future of farming," Stone said. "That's how it's going to become more efficient and better than it is today. That said, we still manufacture these very large, very sophisticated machines. My group is for sure a 100% tech company within Deere."Related posts:
- Oracle's Hurd: 90% of Apps Will Integrate AI in 2 Years
- IBM Cloud Taps Into Nvidia GPUs for AI & HPC
- Workday Acquires Machine Learning Startup SkipFlag
- AI Will Create 2.3 Million Jobs by 2020 – Report