Automation, AI Mark Dawn of Next Farming Revolution

Article By : Anne-Françoise Pelé

Today, autonomous vehicles and artificial intelligence (AI) are marking the dawn of the next agricultural revolution.

Farming is as old as civilization itself and has journeyed through numerous revolutions over the centuries. Today, autonomous vehicles and artificial intelligence (AI) are marking the dawn of the next agricultural revolution. In a spotlight session at CES 2021, Illinois-based machinery producer John Deere described how farmers are embracing the latest technologies to maximize crop yields, profitability, and a sustainable business model.

CES has become the transportation exhibition that sets the tone for the new year. OEMs and carmakers showcase innovations for electric, self-driving, and advanced driver-assistance systems, and for the third consecutive year, John Deere has rallied the community with AI-powered combine harvesters and sensor-fueled planters.

“Every day, as farmers go to the field across the world, they are leveraging technologies like GPS, connectivity, computer vision, and machine learning to make sure they can sustainably grow the food that everybody eats,” said Deanna Kovar, vice president of production and precision agriculture production at John Deere, in the CES session.

Operating autonomously

The concept of fully autonomous agricultural vehicles is far from new. “We’ve had self-driving tractors for almost 20 years,” said Jahmy Hindman, chief technology officer for John Deere. “It’s really important, especially for row crops, to plant the rows in a very straight way to make sure we have equal distance between rows.”

GPS-enabled self-driving tractors and self-propelled equipment confer an additional level of accuracy to the farming operation, and John Deere said it augments that GPS signal with a real-time kinematic (RTK) system that provides pass-to-pass accuracy of ±1 inch. “Our GPS technologies have the ability to get us to within 2.5 cm of accuracy,” whereas cellphones get to within roughly 5 meters, said Kovar.

John Deere's Harvester
(Image source: John Deere)

John Deere acquired NavCom Technologies back in 1999. It now has its own satellite correction system and ground-based correction sites all over the world “to make sure GPS positioning receivers can not only get the planter row unit within 2.5 cm of accuracy but can also come back weeks or months later and get to the exact same spot, all through satellite correction,” said Kovar.

With self-driving cars, the driver shifts over to the passenger’s seat. With self-driving tractors, John Deere aims to shift some of the detail work from the farmer to the machinery, freeing farmers to concentrate on their business priorities. Automated farm equipment is also about making the whole farming operation more effective and productive. With John Deere’s AutoTrack steering system, the farmer can enter the field, get started, and never touch the steering wheel the entire time the field is being planted, according to Kovar.

John Deere’s 16-row 1775NT planter, presented at CES 2021, is a “robot with a ton of automation, leveraging GPS to make sure we know where every seed is placed,” she added. Equipped with 300 sensors and 140 controllers, the planter is claimed to plant 100 seeds per second in a corn operation or a couple of hundred seeds in a soybean operation at 10 mph. “Eighty years ago, a farmer would have planted 30 acres a day; they can now plant 500,” Kovar said.

Hindman noted that every row unit is controlled electrically “to make sure we can precisely grab the seed and then deliver it to the ground in a very precise way. Then we need to make sure it’s at the right depth within the soil. If it’s too deep, it may take too long to emerge. If it’s too shallow, it may not have as much moisture as it needs to germinate. We have to positionally locate that seed very accurately and do that across 36,000 seeds per acre in a typical corn operation or 80,000 seeds per acre in a soybean operation.”

John Deere’s 16-row 1775NT planter (Image source: John Deere)

Precision planting at high speed and on a large scale is essential to have uniform emergence of the crop so that all of the seedlings compete equally for sunlight and for the nutrients and moisture in the ground. The planter also controls downforce, or the amount of pressure exerted on the soil, to firm the ground around the seed and provide optimal soil-to-seed contact. “The end result of that equal competition for farmers is maximizing their productivity, maximizing their yield per plant,” said Hindman.

Self-driving agricultural machinery is already here, but will autonomous tractors take the farmer out of the fields? Absolutely, answered Hindman, and “that future is much closer than it is far away.”

In agriculture, self-driving is more than getting from Point A to Point B and driving the tractor in a straight line. There is a lot of activity going on behind the vehicle, and “we have to make sure that we automate the planter functions appropriately and autonomously,” Hindman said. “The next step in the agricultural process is tillage. It’s a complicated system, and we need to fully automate all of those functions before we can go completely autonomous. But that’s very close and very near-term. I am confident I will see it in my lifetime.”

Connecting to 5G

The advent of autonomous tractors is largely dependent on the broad availability of secure and reliable wireless connectivity. Today, much of John Deere’s equipment is connected through 4G and LTE cellular networks to the cloud, into what the company calls the John Deere Operation Center. “Connectivity is really important, and the amount of data that we push into the cloud in any given growing season gets up to 100 megabytes per second,” said Hindman.

5G will be central to precision agriculture, as it promises ultra-fast speeds and real-time responses. “The lower-latency benefit opens some opportunities in machine control and automation that are difficult to do with higher latency,” said Hindman. “The higher-bandwidth capability also gives an opportunity to start exploring compute at the edge, so that most of the compute is happening onboard the machines.”

Farming is a 24-7 job, and it’s critical to make timely decisions. If farmers miss the perfect planting window in their geographic area, the result is percentage loss in their yield. Leveraging 5G capabilities will help farmers get the data to and from the fields more accurately, quickly, and efficiently. “Farmers have been collecting data [from equipment with embedded intelligence] for over 20 years,” Kovar noted. “They started monitoring and measuring the yield per acre of land and got to understand the relative productivity of different parts of their fields.” How well is the planter running? How many seeds has it placed in the ground? How to optimize agricultural inputs?

Today, data flows to the cloud-based system, and farmers can analyze it on the go via a mobile app on their phone. “No matter where farmers are, they can see what’s happening with their fleet, all the inputs and outputs of their operation, right in the palm of their hand,” said Kovar.

There is a tremendous amount of variability and unpredictability in farming. Weather patterns are a moving target, and there might be three or four different soil types within a single field. Farmers need to have their data at their fingertips so that they can coordinate their plan of attack based on the history of their farm and their ability to interpret it.

Farmers have control over their data, so they can determine who has access to it (for example, agronomists, equipment dealers, seed companies, and field technicians). “We are committed to making sure their data is secure and that farmers are always in control of what’s happening with their data,” said Kovar. “Over 184 connected software companies are using APIs to push and pull data from the John Deere Operation Center any time the farmer allows it.”

Capturing value from AI

AI is beginning to deliver on its promise to provide real value, driven by recent advances in pattern-recognition algorithms and higher computational resources. “Historically, we’ve been data-rich and insight-poor, but we’re quickly getting to the point where we can be data-rich and insight-rich,” said Hindman. “It’s a really interesting time, as you’ve got advances in computational capability coming into play; you’ve got advances in connectivity — 5G being one example, and satellite connectivity being another one; and these advanced algorithms in the AI space that are making it possible.”

In the past decade, John Deere has steadily increased investments in AI technologies to implement its vision for the autonomous farm. In 2017, it acquired Silicon Valley-based computer vision startup Blue River Technology, which uses AI to interpret images captured by cameras installed in machinery and enable autonomous decision-making.

At CES 2020, John Deere featured its AI-enabled R4038 sprayer. Based on Blue River’s image-recognition technology, the smart sprayer can “discriminate between friend and foe in the field, weeds versus crops that we want to preserve,” said Hindman. “That’s a great benefit to growers who have historically done a broadcast spraying operation … and a great example of where vision [technology] is coming into play, coupled with a convolutional neural network.”

At this year’s CES, John Deere described how its new X9 combine series uses multiple AI-based systems to help farmers harvest crops more efficiently. “We’re using AI to monitor the grain flow through the combine, so we’re taking images of the grain through the combine and adjusting settings automatically to maximize the amount of clean grain that goes into the grain tank and separate out the chaff or the things that you don’t want out of the back of the combine,” said Hindman. The X9 combine is equipped with video cameras, and AI algorithms analyze the images.

John Deere - Farming
John Deere’s X9 combine series (Image source: John Deere)

AI is also used for predictive maintenance. The information collected and analyzed aboard the harvester or the planter helps to predict when a failure might occur. “It gives the owners of the equipment forewarning when something might happen so they can take care of it, as opposed to having it impact their business in, for example, the 10-day window when they need to plant,” said Hindman.

AI-empowered sensors are changing the way farmers plant, spray, and harvest. Today, John Deere is using a vast array of sensors, from the seed sensor on the planter that counts the seeds and checks how they are going into the ground to the near-infrared sensor on the harvester that assesses the nutrient value of the food for animals.

The company said it expects to enrich the sensor suite by investing predominantly in camera technologies, computer vision, and machine learning.

This article was originally published on EE Times Europe.

Anne-Françoise Pelé is editor-in-chief of and EE Times Europe.

Leave a comment