Kyler Laird is about to carve a notch into agriculture history by planting 10,000 acres of soybeans with a ghost behind the wheel. The Indiana farmer planted all his corn with a driverless, automated tractor in 2017, punching holes across 535 acres, and he intends to go big in 2019. Skeptics beware: Laird has loosed the DIY robots and his operation will never be the same.
How does a farmer automate fieldwork prior to harvest? Start with a John Deere 420 lawn tractor and climb the driverless ladder—Massey Ferguson 2745, Challenger MT765, and John Deere 6330. Growers often cast dubious eyes at promises of automation, but Laird’s ingenuity is proof in the dirt of technological change on the near horizon. Under the umbrella of his fledgling company, Sabanto, Laird aims to barnstorm across 10,000 acres of farmland from Texas to Canada in a benchmark planting demonstration of equipment utilization and robot efficiency.
Hanging at the Shack
Halfway between Chicago and Indianapolis, Laird, 49, grows 1,700 acres (Lairdscape) of no till corn and soybeans on light, sandy soil spread across relatively flat ground outside the small town of Rensselaer. An understanding of the remarkable innovation spilling from his rows requires a look backward over Laird’s shoulder, at a circuitous route home for a northwest Indiana farm boy.
A young kid of the 1970s, Laird grew up with the advent of the home-gaming industry, but his interests went far deeper than the joystick promises of Atari or Magnavox Odyssey. Whenever his parents made the 45-minute drive to shop in Lafayette, they dropped Laird at his destination of choice—Radio Shack. “That’s where I hung out while they shopped,” he recalls. “Computers fascinated me and it was something that naturally drew me in.”
On Christmas day of 1980, Laird unwrapped a TRS-80 Color Computer, a rarity for a fifth-grader, considering the $399 sticker price. Laird’s father, Wally, took note of his son’s sharpening computer skills and quickly followed the TRS-80 with an IBM PC, tagged at over $5,000. With tremendous foresight, Wally made a stout investment in Laird’s future. “I still have the receipt. It had single-sided floppies and a cassette port,” Laird recalls. “It was quite a price, but my dad was all about anything for education. He totally supported and backed me.”
Whether on the computer or driving the tractor, Laird constantly felt the pull of automation. “Even when I was out disking in middle school, I’d be thinking, ‘This can be automated.’ I was the only one at my entire school with a computer and that’s the direction my mind always went.”
Following a computer science degree at Purdue University, Laird worked for the Purdue School of Engineering for a decade and gained a master’s degree in ag systems management. Beginning in 2001, he helped found the University of California Merced. In 2010, Wally passed due to a brain tumor and Laird’s trip on the farming roundabout came full-circle. “I went home. Agronomy was not my strong suit and I had to figure out what I was good at. I knew automation and I had an idea of how to get there.”
Laird bought several pieces of new John Deere equipment and forged a plan: Pivot away from automation for a five-year chunk in order to learn standard machinery operation and then transition to autonomy. “I wanted to first get a feel for things. My dad hated grain carts and used a six-row combine so I had to get current equipment. I was working and thinking, ‘This can get so much better if I take control of my machines.’”
Cranking the Ghost Mower
In 2015, Laird began dabbling with a driverless John Deere 420 lawn tractor intending to start tiny and go big. The lawn tractor, dubbed Tractobot00, afforded Laird little space for work, essentially forcing Laird to rip out the entire electrical system and hydraulics—all the hoses and hard lines. Counterintuitively, the lawn tractor would prove harder to automate than big tractors. “I had to make my own hard lines, with all the flaring and bending. It was quite an experience and I had to do it from scratch.”
Next up? A Massey Ferguson 2745 tractor in 2016: Tractobot01. “This was comparatively easy. I had plenty of room to work. My goal was simple: How dumb can I make it?”
Laird utilized a bang-bang valve for the steering, a ride height sensor to indicate how far the front wheels turned, and a GPS receiver to monitor location. He used a simple, linear actuator on the clutch and another bang-bang valve to raise or lower the implement: “I logged in to the tractor with my laptop with text interface and drilled 50 acres of beans. It did a much better job than I ever did and I learned great lessons.”
Also in 2016, Laird introduced Tractobot02, an MT765 Challenger he used to pull a grain cart during harvest. (In 2017, he also used Tractobot02 for vertical tillage and at planting in 2018.) The Massey Ferguson automation had required a collection of physical controls, but the MT765 only needed electronic signals to mimic steering, throttle, transmission and hydraulics. Laird pulsed separate lines at different frequencies, essentially turning them on and off. The automated system used an Arduino for low level controls, and a Rasberry Pi as the central brain. “It was fairly trivial to control. I just had to wire everything and go.”
Laird’s spring 2017 project was Tractobot03: a John Deere 6330 with a Precision Planting-laden 8-row John Deere 7300 planter. He planted 535 acres of corn with Tractobot03 and it won the 2017 agBOT Seeding Challenge. “I used a linear actuator for the reverser, and an off-the-shelf standard guidance hydraulic valve which I had someone put on for me,” Laird describes. “Then I wired in to control the throttle and the hitch; that was it.”
Small Steps Forward
The cost of the components differed with each vehicle, ranging from $100 to several thousand dollars, plus hundreds of hours of labor. However, over the long-term, the savings are tremendous, according to Laird: “I can’t afford to hire people; full-time employees are not an option. I’ve never tallied what this saves me now, but I know what it costs to hire someone who can go all the time, day in and day out.”
Rather than hire labor, Laird says the economical solution on his operation is to buy open-station small tractors and automate for a few thousand dollars: “That puts me ahead. I don’t know the savings, but to keep on spending money is not an option.”
Currently, Laird hasn’t invested in high-dollar obstacle avoidance technology, but he believes advances in the automobile industry will find a solution and lower prices. Laird also acknowledges the shortcomings in current automation systems. In 2017, while Laird was combining, his grain cart tractor went too far north and into a neighbor’s field: “I sold the combine and decided to concentrate on the tasks at hand.”
Laird receives frequent questions and encourages interested growers to visit Robot Operating System Agriculture (ROS-A), an open source online community centered on robotic tools. He points to a steep learning curve, but says small steps are the best encouragement for other farmers: “Just start with an actuator to steer, or to raise and lower the gate on an auger. Stick with small tasks and you’ll figure it out.”
Laird is currently working on a robot soil sampler, and consistently fine-tunes the tractobots. “I haven’t run into any patent issues with equipment manufacturers. I bypass the software and don’t tweak it; I’m just pretending to be another operator.”
Tomorrow’s Farming World?
What’s next for Laird? 10,000 acres of soybeans. Alongside Sabanto business partner Craig Rupp, co-founder of 640 Labs, Laird aims to plant across the U.S. and into Canada during 2019, with a single, driverless tractor pulling an 18-row, 30’ planter.
The Sabanto duo will select a new, off-the-shelf tractor model and apply the necessary automation modifications. One single unit will plant the acreage as part of a roving, unprecedented demonstration. “This is a beginning step. It’ll be an inefficient demo and have lots of support, but this is also a step towards something we plan to scale up,” Laird emphasizes.
“Machinery is underutilized and this is our attempt to show how we can better take advantage of equipment,” he continues. “For years I used a 40’ planter on 1,700 acres; now we’re going smaller with a 30’ planter, yet covering 10,000 acres. It’s a first move to show how automation works and how we’ve solved some efficiency problems.”
Laird’s innovations are highly significant, according to Mark Young, CTO of The Climate Corporation: “What's really interesting is what one guy in his garage has managed to do. Now, imagine what ag companies with investments in R&D will be able to do in the next few years. Automation in agriculture is much closer than what people think.”
Kim DeWees, owner of Vision Ag Farm Management Services LLC, in Rensselaer, says Laird consistently challenges agriculture’s status quo. “He challenges the way things are done, and he’s always looking for a better, more efficient way. People are taking notice of what Kyler is doing. They see machines driving around without people in them and it makes them pay attention. He fixes bugs as they arise, improves each machine as time goes on and people are taking note.”
“Kyler is very intelligent, an out-of-the-box thinker. When it comes to technology he can computer program a machine to do what many can only imagine. I think what he’s doing is going to be a big part of our future and we’re just getting a small glimpse of it right now,” DeWees adds.
On Laird’s Jasper County operation, tomorrow has arrived early, but like most farmers, Laird’s shed and shop are filled with millions of dollars in equipment, and the irony grates on his constant drive for efficiency. “I use my expensive equipment like other farmers use theirs—not very often. It’s a horrible, horrible utilization in farming. I see automation changing all of that. In just a few years, it’s going to be a much different farming world.”
See here for a full collection of Laird’s tractobot videos.
For more articles, see: