How ag machines use GNSS for greater harvesting efficiencies

Chris and Cam Clifford, owners of Agri Automation, which distributes agricultural robots. (Photo: Chris Clifford, Agri Automation)

An early 1900s Italian folk song tells of a farmer walking into his fields at dawn to spread wheat seeds with his hand from a small bag.¹ Farming has changed quite a bit since then. After remaining essentially unchanged for about 12 millennia, in the past century, it has been transformed by such innovations as tractors, electrification, chemical fertilizers and pesticides. In the 1990s, precision agriculture (PA) emerged. (This magazine produced a few supplements on the subject around 1999. If you still have any of them, please let me know.)

PA reduces inputs of water, fertilizer, seeds, pesticides and fuel and increases harvests by mapping variations in soil characteristics and plant health and then using those maps to adjust the inputs using variable rate technology on sprayers. It also ensures that no part of a field is sprayed twice or missed and greatly reduces overlap in seeding and tilling. Double spraying is costly and wasteful; missing a row when spraying pesticide can cause pests to concentrate there and then spread, nullifying a whole spraying operation.

The data for the maps are gathered from sensors on tractors and other farm machinery in the fields, as well as by aerial platforms — nowadays, mostly UAVs. GNSS receivers are essential in guiding the farm machinery. The required accuracy depends on the crop but is typically at the centimeter to decimeter level.

Increasingly, farm machinery also incorporates a variety of other sensors, both to compensate for GNSS outages and to minimize the risk of collisions, such as when a cow crosses the path of a tractor. To maintain navigation during GNSS outages, inertial navigation is used. For obstacle avoidance lidar, radar and stereo vision cameras are used to measure the distance to the object. (Both challenges — navigation in GNSS-denied areas and obstacle avoidance — and their solutions are very similar to those encountered with autonomous vehicles on roads.)

In-cab displays enable growers to monitor their progress in real time. They often also download the data and maps to a laptop to better identify missed spots or areas with special issues and to plan their next task.

Manufacturers of PA equipment compete in a global market. Some challenges are the same everywhere, while some are specific — such as strong ionospheric scintillations in Brazil or antiquated agricultural practices in Japan’s Furano region. For this year’s cover story on PA, I discussed these challenges and the latest generation of farming hardware, software and services with

■ Kirstin Schauble, director of systems engineering, ANELLO Photonics

■ Joey Koebelen, founder and CEO, Deep Sand Technology

■ Chad Huedepohl, PA portfolio manager, autonomy and positioning division, Hexagon

■ Ken MacLeod, director of product management and Gordon Echlin, director of business development, Calian GNSS.

This article contains a few excerpts from those interviews. I also received case studies from AgLeader Technology, ComNav Technology and Harxon Corporation.

ANELLO Photonics makes silicon photonics optical gyroscopes, which enable accurate dead reckoning without GNSS and are targeted mostly at the autonomy market. (Anello means ring in Italian, which reflects the nature of the company’s technology and the Italian-American background of its CEO, Mario Paniccia.) Because ANELLO specializes in high precision in situations with obstructed GNSS signals, orchard cultivation is one of the agricultural practices in which it specializes. “Orchards have high-value crops, such as almonds or walnuts, and you’re driving your tractor between very narrow rows with trees completely covering the sky above you,” Schauble said. “Our job is to replace that GNSS input with our inertial navigation system (INS) input.”

Deep Sand Technology — in partnership with GEODNET, the largest real-time kinematic (RTK) network in the world — sells affordable RTK corrections to farmers. It also maintains and troubleshoots the system, compensated by the network’s cryptocurrency. “We handle the blockchain and use it for maintenance,” said Koebelen. “We have someone that checks every day and makes sure that the bases are up. We do the support on it. Instead of charging for that, we take the tokens; that’s just our part of the program, and they get free RTK.” Koebelen, who is also a peanut farmer, adds: “You can trust anything that we sell because it has been tested and used by a farmer and is supported by a farmer.”

Hexagon, a very large company, makes a wide range of sensors that capture and display data about physical reality. Its latest contributions to PA include the TerraStar-C PRO and the TerraStar-X Corn Belt corrections services, which incorporate improvements in ionospheric resiliency. “Especially in the Brazil market, some growers were often experiencing hours of downtime due to ionospheric scintillation,” said Huedepohl. “With the ionospheric enhancements that we’ve added, that downtime now is down to just a few minutes here and there.” He also cites safety enhancements for the autonomy market, such as dual antenna solutions and geofencing.

Calian GNSS is a global supplier of technical solutions, services and products to the space communications, defense, wired and terrestrial wireless, manufacturing, GNSS, agricultural technology and nuclear industries. The company’s recent entries in the PA market include GNSS antennas with lower elevation gain and extended filtering. “Our GNSS agriculture antennas support centimeter level precision, have best in class lower elevation angle gain enabling L-Band correction reception (at northern and southern latitudes), and have eXtended Filtering (XF), which creates very deep attenuation of nearby out of band radio frequency signals,” said Echlin. “Having a digital signal from the antenna to the smart ag controller simplifies and reduces the cost of the installation,” said MacLeod.

Ag Leader, founded in 1992 and focused exclusively on precision farming technology, offers a complete line of systems that integrate with existing farm machinery. In February, it introduced the RightPath passive implement steering solution to alleviate the problem of trailed implements drifting off the guidance line by up to 10 inches or more, even when farmers utilize auto steer and on flat ground. RightPath keeps implements centered on the guidance line, ensuring precise input placement and increasing operational efficiency throughout the growing season while minimizing crop damage, yield loss and operator challenges, Ag Leader said. To utilize RightPath, both the vehicle and the implement require Ag Leader’s GPS 7500, but only the vehicle needs to be equipped with TerraStar-C, TerraStar-X, or RTK. RightPath will be available in late fall 2025 through a single purchase unlock and without any recurring subscription fee.

ComNav Technology is an original equipment manufacturer (OEM) that develops and manufactures GNSS OEM boards, receivers and solutions for high-precision positioning applications worldwide. Japan’s Furano region is renowned for its vast farmland and abundant agricultural resources. Still, it is challenged by traditional manual driving methods that provide insufficient accuracy, low efficiency, and operator fatigue during prolonged tasks. To address these issues, ComNav introduced the AG502 autosteer system, which integrates satellite reception, positioning, navigation and autonomous driving. It is compatible with a variety of mainstream tractors on the market and is suitable for a wide range of agricultural tasks such as ridging, seeding, spraying and harvesting, ComNav said. In the Furano project, the AG502 demonstrated its versatility through its successful deployment on a transplanting machine.

Harxon Corporation makes GNSS positioning antenna solutions. The company has been collaborating with Brazilian agricultural navigation solutions and systems developer Agres to integrate Harxon’s Smart Antenna into the AgresAutopilot System. This secure and robust agricultural navigation solution has been widely adopted by Brazilian agribusinesses to provide automatic steering on straight or curved parallel lines to assist with such field operations as preparing the soil, planting seeds, cultivating the plantation and harvesting the crops. These systems are suitable for various brands and tractor/vehicle models such as Kuhn, John Deere, Valtra, Massey Ferguson, New Holland, LS, Landini, Jacto and others. 

Challenges

The key technical challenges faced by PA systems include minimizing multipath and RF interference and monitoring the positions of implements relative to the tractor. “Agriculture requires positional accuracy, so mounting an antenna on a farm machine is not a trivial matter,” said MacLeod. “On metallic machinery, radio frequency surface currents and reflections (multipath) will degrade the antenna radiation pattern, and RF noise coming from other electronics on the machine can interfere with GNSS.” Additionally, because most GNSS applications now are full band, “the challenge is designing antennas that are small and full band, and which also reduce local multipath on the machine.”

Regarding the position of the implements, MacLeod said: “Many agricultural applications use the moving base technique to estimate a precise heading which can be used to monitor pass to pass overlap. Calian GNSS have smart antennas that support the moving base application.”

Accuracy and Reliability

Nearly all PA practices require RTK, which gives repeatable accuracy of 1 cm to 2 cm. “I have this conversation daily with farmers,” Koebelen said. “All crops or farm practices benefit from RTK, even if you’re just doing hay work — whether you’re planting or harvesting. We can’t control the weather, commodity prices, or fuel prices but we can reduce input costs. So even if you’re just tilling, GEODNET RTK will pay for itself and is better than using traditional autosteer, because you’re eliminating all overlap.”

Additionally, farmers need reliable repeatability, even from one season to the next, to be able to return to the same spot to harvest what they planted. “Peanut farmers may plant with RTX or SF3, but satellite-based corrections, even higher precision ones, didn’t provide them enough repeatability to come back to harvest,” said Koebelen. “So, they still had to adjust their lines or hand-drive them. If the spacing between passes are off by even two to three inches, you’re going to lose peanuts. That’s why peanut farmers — as well as growers of potatoes, cucumbers, and other crops — need RTK.”

Once they enter a GNSS-denied area, such as an orchard, farm machines will need a dead reckoning capability that can keep them within a 20 cm to 30 cm error, said Schauble. “This is typically posed as a cross-track error. Errors in the direction of the distance traveled are slightly less important, because you can tell based on visuals when you exit a row.”

Growers think of reliability, accuracy and repeatability in terms of whether they can count on a system to do what they are asking it to do, Huedepohl explained. “They think about all those things. They do not necessarily focus on one thing versus another.”

Retrofitting

While many agricultural systems are proprietary, there is also a lot of mixing-and-matching and retrofitting going on. More than 90% of new tractors come with factory-installed guidance, but some growers want to retrofit new receivers on their machines, either because they did not have them or to upgrade. On some machines, it is possible to feed better positioning data — for example, integrating GNSS and inertial navigation — into the port that previously took in only GNSS data, using a standard NMEA format.

“It’s a simple plug-and-play to exchange someone’s GNSS receiver with our INS solution. Obviously, they need to do some testing to optimize placement, installation and stuff like that,” said Schauble. “Many companies are retrofitting existing tractors with an autonomy stack. They take commercial off-the-shelf (COTS) systems, such as ours, or a lidar or a camera, and retrofit a tractor. That’s their business model.”

Additional Sensors

Among the additional sensors often used are wheel odometers. “Without the wheel speed, you’re relying heavily on accelerometers,” said Schauble. “Growers cannot afford to pay $100,000 for a reference-grade system. The navigation systems for these applications use MEMS accelerometers, as we do. So, wheel speed aiding is extremely important to maintain that distance traveled.”

Integrating GNSS and inertial measurement units (IMUs) has long been standard. Increasingly, this integration is done inside an antenna, called a smart antenna. Calian, among others, does that. “We also have smart antennas that employ the L1-L5 observation pair rather than L1-L2, since the L5 signal is stronger and performs better under cover,” said MacLeod. “L5 uses an enhanced signal architecture with 10x faster chipping rate (10.23 MHz) offering more precise standard localization and improved multipath mitigation for reflections exceeding 29.3 m.”

Corrections

Corrections have also been key to the evolution of PA. A reference base station can provide 1-inch accuracy for up to 21 miles, degrading beyond that distance. It needs a WiFi network to communicate, so farmers often place the base station near their home and connect it to their home network. “We haven’t found an internet connection that isn’t quick enough to handle that,” said Koebelen. “From there, you can use your hotspot with a SIM card on your phone, and it’s like texting, so it will not drop like with voice calls. We haven’t run across rural areas where cell coverage is the limiting factor.”

RTK adoption is growing among farmers. “In the past, many people did not want to use RTK, because it was not very affordable nor easy,” Koebelen said. “However, now that we have these networks [such as GEODNET], you’re going to see a lot more people rely on the precision of RTK and you’re going to see many new products come out. Right now, even John Deere, Trimble and other major brands that are more expensive are trying to make the tier below RTK more affordable or easier to get — for example, RTX, SF3, the satellite-based corrections.” GEODNET’s network is growing rapidly, he said, “because our price for RTK is lower than Trimble’s or John Deere’s basic entries, which use free satellite signals that drift throughout the day.”

Huedepohl agrees that RTK has improved while prices have dropped significantly. “Earlier in my career,” he said, “RTK positioning was very expensive and satellite-based augmentation systems (SBAS) were not as stable. Also, RTK systems and such used a single constellation for the longest time. We started adding in GLONASS and then positioning network (Ntrip) corrections, which gave us a lot more robustness.”

Precise point positioning (PPP) has also improved. It used to have convergence times of up to 45 minutes. “Then, you would drive underneath one tree on the edge of a field, and you had to start all over,” Huedepohl recalls. “That did not sit well with farmers, so PPP corrections struggled to take off. Because of those early experiences, it took a long time for the market to start to accept the newer PPP models that we’ve seen in the past seven or eight years. Now there are farmers who enjoy the reliability of those PPP corrections.” The convergence time for one of Hexagon’s PPP services, TerraStar-C PRO, is often less than five minutes, according to Huedepohl. “We have a fast startup time. So, if the tractor was shut down, already converged and you turn it back on, most people are going to be reconverged in just a minute or two.”

Division of Labor

The division of labor between manufacturers of PA equipment depends, in part, on whether a system is a retrofit or built from scratch. “If you are, let’s say, John Deere, and you own the entire autonomy stack within this tractor, then you can take our INS solution, add cameras, maybe add a lidar, and you can have your own fusion of those sensors,” said Schauble. “We have our own sensor fusion with IMUs and GPS. The tractor’s autonomy stack can do the sensor fusion with our output and other visual sensors, such as cameras and lidars.”

“Dealerships want their tractors to be known as having the highest tech,” said Schauble. “For a dealership to offer our state of the art, autonomy-enabling technology would be a huge benefit to them.”

Another differentiator is whether a factory-installed system is an OEM or branded. “We’ve been providing NovAtel branded receivers to AGCO for many years, through their channel, both factory-installed and aftermarket. Some of the others, such as CNH, are white labeled, so it would just say ‘Case-IH’ or ‘New Holland’ and have no Hexagon markings.”

Whether OEM or aftermarket, most manufacturers have some type of proprietary integration. “There are products that are just NEMA; they are typically at the lower end and priced much lower,” said Huedepohl. “The higher performing flagship products out of everybody’s portfolio are usually doing a more customized integration.”

Echlin has a similar perspective: “We provide products to OEMs who designed our products into their machinery. There are also system integrators and aftermarket system providers that use our smart antennas.”

According to Harxon, one reason for the success of its smart antenna in the agriculture market, especially for autonomy users, has been its ease of integration and high performance. “GNSS positioning is just one part of an autonomous system, and the autonomous integrators don’t necessarily have resources or expertise to develop an OEM component portfolio. Therefore, it’s a timesaving and cost-effective choice to directly integrate a smart antenna into an autonomous system.” 

¹ “Di buon mattino il contadino va nei suoi campi a seminare il grano. Ha un sacchettino e ci tuffa la mano.”