Precision agriculture is a unique balance of advanced technology and human expertise. Farmers know their land better than anyone else, but they face a big challenge they can’t solve on their own: feeding our growing population with less available land, labor, and resources.
Experts predict that employment of agricultural workers will only grow 1% by 2029, but global food demand could double by 2050. How will farmers meet this demand despite a shrinking labor pool and limited resources?
According to analyst firm Gartner, “Making more effective business decisions requires executive leaders to know when and why to complement the best of human decision making with the power of data and analytics and AI.” Farmers are the executive leaders of their farms, using data and advanced technology to make vital decisions about the thousands of microenvironments in their fields.
It may come as a surprise, but farmers have always been tech-savvy and early adopters of new technology. On the average farm, you’re likely to find as much technology as in Silicon Valley. Farmers routinely use GPS, machine learning, self-steering vehicles, and, most recently, fully autonomous machines.
Think like a farmer: training vision systems with data and machine learning
At the highest level, fully autonomous machines must do two things: perceive the environment around them and make intelligent decisions about what they perceive. Autonomous tractors may use stereo cameras to see and understand their surroundings. Cameras act like human eyes for the tractor, providing a 360-degree view as it moves through the field working and checking for obstacles.
But the tractor needs a brain to decide what to do next. Machine learning algorithms running on GPUs interpret the images from the cameras and perceive the world around the tractor. If the machine learning model picks up anything that isn’t the ground or sky, it stops and alerts the farmer. The farmer is then able to make a decision on whether to navigate around the obstacle or to physically clear the obstacle.
Even though farms have fewer variables than open roads, perception systems for autonomous robotic tractors still need to be well-trained to ensure they perform tasks precisely and safely. These tractors must interpret their surroundings and monitor for potential hazards, all while performing time-sensitive and complex tasks. The vision system monitors for broader areas like ground, trees, and skyline while also accounting for obstacles, such as large rocks, animals, or debris.
Move like a farmer: accurate and precise movement with a purpose
Farmers instinctively know how to move throughout their fields. They know every curve, bump and rough patch. But, autonomous tractors use GPS and positioning data—which first appeared on precision agriculture machines in 1997—to move more precisely. Using an advanced GPS system—accurate down to less than an inch—farmers can map a dynamic path for the autonomous machine to follow. This digital “fence” helps the tractor know its work zone and the specific route it should take to get the job done.
Unlike autonomous vehicles on the road, autonomous machines on the farm need to do more than just move from point A to point B. They must move from one end of the field to the other while performing multiple complex tasks with extreme precision, including tilling, planting, and spraying. For over 20 years, precision agriculture machines have been self-steering—but an operator still needed to stay in the tractor cab. But now, because of precise positioning and movement prescription, farmers can step away from their field confident in the machine’s ability to get the job done right the first time.
Act like a farmer: the fully autonomous farm of the future
Every new autonomous advancement must prioritize farms and farmers. By focusing on technology for the most time-consuming and demanding parts of the growing cycle, autonomous precision agriculture equipment can replicate a farmer’s efforts. This lets farmers focus on other priority jobs, have time to manage their business, or simply spend time with their loved ones. Deploying autonomous technology could make the difference between a farmer working through the evening or getting home to enjoy dinner with their family.
As technology continues to advance, machines will perform more and more complex tasks autonomously. The industry continues to build out training data, sensing and vision technology, and machine learning and AI algorithms, in order to find more ways to integrate this advanced technology with farmers’ knowledge. The result is a more precise and successful operation to feed a hungry world.
Gaurav Bansal, PhD., is director of engineering and autonomy at Blue River Technology, a John Deere Company, where he leads an engineering team delivering the machine-learning stack for a new autonomy product.
Editor's Note: To learn more, attend Gaurav’s conference session, “The Future on the Farm is Now - The Technology Behind Autonomy,” at Sensors Converge on June 28 at 1:50 pm PT. Register for the event here.