The University of Nebraska-Lincoln's Eastern Nebraska Research and Extension Center is well-known as one of the only places in the U.S., if not the world, with a Spidercam used for collecting crop phenotype information in the field.
The Field Phenotyping Facility, located near Mead, Neb., is home to four 90-foot-tall poles, each connected to a cable that helps hoist a dolly carrying a Spidercam along with other sensors used to collect phenotypic data on corn and soybeans in the 1-acre field.
This summer, there was a new piece of technology traversing the 1-acre field — the Flex-Ro, short for flexible robot, built by UNL's Biological Systems Engineering Department.
The Flex-Ro, which can be controlled remotely and is capable of full autonomy, looks almost nothing like any machine that's traversed the corn and soybean fields nearby. It sits on an adjustable high-clearance platform, weighs about 3,800 pounds and is powered by a 60-hp Kubota gas engine.
It uses a hydrostatic transmission, using a hydraulic pump powered by the engine that's controlled through a Controller Area Network, or CAN-bus system. Each wheel has an electric and hydraulic motor. This way, each wheel can be programmed and controlled individually, with different steering modes — four-wheel (crab) steering, differential steering (similar to a bulldozer), and normal tractor steering.
First time in the field
This is the first year it's been in the field, however. Santosh Pitla, associate professor of advanced machinery systems at UNL, who has led the Flex-Ro project since its inception a few years ago, notes the goal this year is to collect phenotype data through onboard sensors such as an RGB (red, green, blue) camera, a spectrometer to measure reflectance, temperature and humidity sensors, and ultrasonic sensors to measure height, while comparing with measurements from the Spidercam.
"We're going to establish a correlation between the two, because the Spidercam is considered the standard, and we'll know what we need to do to adjust the Flex-Ro," Pitla says. "We'll have a mobile phenotyping facility, so we can go to any field. It doesn’t have to be 1 acre — it can be 200."
The Flex-Ro has been upgraded in the past year, with a new obstacle detection sensor mounted on the front, a new toolbar to mount sensors, and an auto-guidance system. This year, Pitla also is testing how the machine navigates the field using the new guidance and obstacle avoidance technologies.
Josh Murman, UNL Biological Systems Engineering graduate student, programmed the auto-guidance system and designed and wired the toolbar that carries the sensors. He designed the toolbar to break away, so that it won't break if it collides with an obstacle.
Murman also programmed the controller specifically for the Flex-Ro using components donated by Danfoss Power Solutions. This includes a heads-up display that shows the operating or steering modes, and monitors fluids and movement of the machine itself.
While the controller uses a joystick to control the Flex-Ro, he also programmed the machine to be controlled from a laptop.
"The joystick is intuitive,” Murman says. “I've had tours that come in and people can operate it right away. It's almost like playing a video game. Now, you can also drive with the arrow keys on a laptop, which is pretty convenient, too. And you just hit the space bar to stop."
Vision for the future
The Flex-Ro is intended for multiple purposes and is built to be modular — that is, seed hoppers and sprayer tanks can be added on. Also, an electronic control unit, or ECU, can be added to control different applications.
Pitla hopes to eventually use the Flex-Ro for seeding cover crops by mounting a bulk seed hopper. However, he notes it isn't yet clear what the ideal method for seeding from the Flex-Ro would be.
"Now that we have a machine we can drive through a standing crop, or even attach a one- or two-row planter unit, we're not limited anymore. So what is the ideal situation for planting cover crops?" he asks. "We also want to characterize the benefits of planting cover crops by taking soil samples using the machine to keep track of soil quality and quantify the improvements."
The biggest application likely will be managing resistant weeds. Pitla says this will involve creating algorithms for onboard computing, so the machine can distinguish between weeds and cash crops.
One option for controlling weeds using the Flex-Ro may involve a robotic arm mounted beneath the machine for mechanical weeding and selective spraying. However, different kinds of robotic arms also could be used for phenotyping purposes — such as measuring chlorophyll content.
"For me, the vision is this is always in the field,” Pitla says. “That's what AI is. It should have decision-making. Because there's no operator in the machine, it can move slowly, and constantly move through and scout in the field. If there's a weed, let's take care of that."
Pitla also is leading research efforts to advance tractor-testing methods at the Nebraska Tractor Test Laboratory. With greater interest among growers and manufacturers in autonomous equipment, Pitla notes one of the steps in the future will be testing autonomous equipment.
"We're in a similar position as the industry was 100 years ago,” he says. “At the time, farmers were buying tractors, but there was no benchmarking and no testing to verify manufacturer's claims. In the future, we will be testing for intelligence, not for power. That's where I think we need to be — a next-generation autonomous tractor-testing facility."