The need for affordable, but reliable help is just one reason corn and soybean growers will likely turn to robots in the near future to accomplish tasks that humans do now, says Tony Grift, University of Illinois agricultural and biological engineer.

The first use for row-crop robots will likely be scouting, says Grift. “If you put a simple (video) camera on a robot, you can already get good data now,” he explains. “Using a camera, weeds can be recognized quickly, and with global positioning systems (GPS), we can do weed maps.”

With inexpensive robots regularly scouting fields, farmers can stay informed about crop conditions and act quickly to solve problems, says Grift. All they need to do is take note of the GPS coordinates on the video and make spot treatments where needed.

However, Grift's vision for robot farming goes well beyond Star Wars-type androids that walk through corn rows with video cameras. He's currently working to create several different robots that hold promise for performing much more valuable farm functions.

“We're developing robots that will cost $150-7,000,” says Grift. “They are a smaller, but smarter approach to farming than current machinery.”

One robot is named the “AgAnt,” and it is the cheapest of the row-crop robot types Grift is developing. Just one foot long, this robot can still adroitly maneuver through corn rows. In the near future, the AgAnt could be used to scout row crops and detect weeds, disease, insects and other crop stressors.

The next step is to develop other robots, with whom the AgAnt can communicate, which will apply pesticides and fertilizer with greater precision than current machinery, says Grift. He adds that the ultimate goal is to replace large, expensive pieces of farm equipment with smaller, less-expensive robots.

“Reliable, no-calibration, real-time sensors are being developed for the detection of stresses, diseases, weeds, pests and varying soils,” points out Grift. “However, we don't have a lot of good sensors yet to do everything that is needed, such as a real-time corn rootworm larvae detection system. Right now, all we have is the camera.”

Soil sensing is a different matter, however, says Grift. “We have good sensors now that can measure real-time moisture levels, compaction levels and pH,” he points out. “What we really need is one that measures real-time nitrogen levels.”

However, useful, on-the-go soil mapping systems have already been developed, agrees Viacheslav Adamchuk, University of Nebraska agricultural engineer. For example, a commercialized electrochemical on-the-go sensing system is now available from Veris Technologies (see “Pay Dirt From pH,” February 2004). Other soil mapping systems are at different stages of development and testing, he adds.

The system from Veris Technologies is called the Mobile Sensor Platform (MSP). It can be pulled behind a pickup truck and features continuous soil electrical conductivity sensing and pH measurements that are automatically taken every 10 seconds. Along with GPS equipment, these sensors help to produce soil maps and allow farmers to variably apply lime and consider other differentiated soil treatments.

“Although the pH sensor doesn't necessarily tell you the lime requirements, it does identify field areas needing lime,” says Adamchuk. “We use it to show the highs and lows on a map. Then we take manual soil tests and use those as a calibration for the map, which helps us to perform variable-rate applications.”

On-the-go soil sensing is typically less expensive and often more accurate than traditional grid soil sampling, says Adamchuk. “Typically, the on-the-go sensor can do 10 times higher the sampling rate at the same or lower cost as a 2.5-acre grid sampling system,” he says. “However, the advantages from its use depend on the field. For fields with lots of variability, the on-the-go system shows higher payback.”

On-the-go measurements from mechanized tools like the MSP have greater potential for use now, compared to autonomous robots, because they fit with current precision farming machinery, says Adamchuk. Some soil sensors currently under development will likely be attached to a toolbar, he adds, and then adapted for use with other field operations.

“So far, using autonomous robots to make soil measurements is somewhat futuristic,” says Adamchuk. “In general, robotics for scouting still has many limitations, including providing them with enough power to operate efficiently for long periods. However, hypothetically, you could mount one of these recently developed sensors onto a robot and if smart enough, the robot could even decide by itself how far away to make the next measurement based on geo-statistical field data.”

The problem with using either robots or on-the-go machinery to perform chemical soil tests is that most applicable ion-selective electrodes were developed for use in slurry or for solutions under controlled conditions. “None of them are designed for the kind of impact that occurs when you insert an electrode into a core of soil,” says Adamchuk. “However, for pH they work fairly well. For nitrate, potassium and some other ions, the reliability of recently developed on-the-go measurements remains in question.”

Still, today's limitations aren't stopping future development efforts, says Adamchuk. “We are already working on a system that maps soil pH, residual nitrate and soluble potassium at the same time.”

Other automated sensors are already providing many valuable, on-farm measurements, points out Adamchuk. “For instance, a blade sensor can now measure soil mechanical resistance, which frequently correlates with soil compaction,” he says. “Those sensors can also show how resistance changes with depth and whether compaction is on the surface or in the soil pan.”

Additional sensors have been developed to measure moisture content and/or soil organic matter. “We've mounted a combination of different sensors to a single toolbar and pull it behind a tractor to obtain several different maps in one pass,” he adds.

Above-ground sensors have also been developed to monitor crop conditions. “Companies such as NTech Industries, Inc., and Holland Scientific, market on-the-go plant sensors in the U.S.,” says Adamchuk. “NTech's product is called GreenSeeker and Holland Scientific's product is called Crop Circle. In Europe, Yara markets a plant sensor, called N Sensor, which has about 400 users right now.”

With these technologies, farmers can change sidedress nitrogen applications based on canopy reflectance. “You do this in mid-growth, when the plant just starts showing signs of nutrient stress that the human eye is not yet able to detect,” says Adamchuk.

Weed sensors are another promising automated technology currently available on the market, he adds. This new technology allows machinery to recognize common weeds and then spray them on the go. Commercially available systems allow weed detection on bare ground, such as along railroad tracks and spray herbicides accordingly.

Robots that can scout, soil test and kill weeds may prove, in the future, to be even more economical than current, automated prescription farming tools, says Adamchuk. “However, in what form robots will appear, and what they can do 10 years from now, remains unclear.”

When combined with remote sensing (satellites or drones), which provides a good overview of the crop from above, robots could help fill in the details on the ground, says Grift. “It just depends on what sensors are available to us. The key is to continue to develop more advanced sensors,” he says.

With the right technology, robot farmhands could monitor fields, both day and night, to detect varying crop and soil conditions that might be causing plants stress and then relay that information back to the farmer's computer or to other robots, says Grift. More specialized robots could then perform various tasks to relieve stress on those plants, such as weeding, watering, spraying or fertilizing. Future robot tasks might also include planting, soil testing and harvesting.

Pocock Potential Uses For Row-Crop Robots

What will row-crop robots do on grain farms of the future? Most likely they won't take over your farm management duties, but they'll work very reliably as farmhands, says Tony Grift, University of Illinois agricultural and biological engineer.

“I doubt robots will ever replace humans completely on the farm,” says Grift. “Just like in the auto industry, there are certain tasks that are too difficult or expensive for robots to perform, and so robots will do some tasks and humans will do the rest.”

Yet, robots will probably be able to perform many farm-related tasks more precisely and efficiently than humans, emphasizes Grift. He says the potential benefits from using robots on future row crop farms are numerous. They include the potential to:

  • apply crop protection products and fertilizer with greater precision and less risk to the environment than is now possible with current machinery,

  • perform field operations around the clock, while farmers sleep or carry out management functions,

  • increase farm safety by acting as guards, or watchdogs, or by having them perform tasks, such as insecticide applications, that might be dangerous for humans and

  • decrease compaction by replacing large machinery with a fleet of inexpensive, lightweight communicating and autonomous machines.