As Drone Navigation Advances, SlantRange Focuses on Farm Analytics

As the Federal Aviation Administration sets new rules for commercial operations of unmanned aircraft, legions of new startups are using drones to perform all kinds of tasks—from aerial inspections to mapping, imaging, and analysis.

Whether these new businesses can succeed will depend on a variety of factors, including restrictions on autonomous flight that government regulators are still trying to formulate. Until then, new FAA rules adopted last month require a certified operator to remotely fly small, unmanned aircraft (weighing 55 pounds or less). These radio-controlled aircraft  must stay within view, or what the FAA calls visual line of sight.

In the meantime, the robotics industry is racing to advance “intelligent navigation” technologies intended to enable unmanned aerial systems to fly autonomously out of view, and to avoid telephone lines, trees, and other obstacles along the way. Intelligent navigation is crucial if you’re in the business of using drones to deliver packages, and drone startups have been trying a variety of approaches to autonomous flight in complex environments. For example, at Xconomy’s Robo Madness forum in San Jose, CA, last month, three companies each presented a different approach to intelligent navigation:

PreNav, founded in San Carlos, CA, in 2013, has developed a separate, ground-based system that uses a pulsed laser (formally known as Light Detection and Ranging, or LIDAR) to precisely determine the flight path for a quadcopter. CEO Nathan Schuett said PreNav uses the technology to conduct inspections of wind turbines and other tall structures by scanning a structure and generating a 3D map that enables the drone to fly with centimeter-level accuracy.

Swift Navigation, founded in 2012 and now based in San Francisco, is using real-time kinematic (RTK) technology to enhance the precision of satellite-based GPS positioning data. Using the company’s GPS module and software, CEO Tim Harris said, “We have to provide error correction that gets [GPS accuracy] from 15 feet to about 2 centimeters.”

5D Robotics, founded in 2009 in Carlsbad, CA, (about 30 miles north of San Diego), has developed a device that integrates GPS, inertial guidance, optical positioning, and other technologies. Because GPS doesn’t work well indoors, underground, or in the skyscraper canyons of big cities, 5D has added land-based beacons that use ultra-wideband (UWB) wireless transmitters to guide aerial drones and ground-based robots. “We do UWB as one more thing in the tool kit,” said CEO David Bruemmer, who recently demonstrated what he calls “virtual tether” technology that enables a drone to land atop a moving vehicle.

Until the regulations governing autonomous flight are settled, though, intelligent navigation remains secondary to a more basic concern—making money.

SlantRange, a San Diego company founded almost three years ago, is a drone-related startup already generating sales. Instead of making drones or offering commercial drone services, SlantRange makes a sensor for drones used in agriculture, and provides specialized analytics for the immense amount of data it generates. Their customers are the agronomists (or the drone service providers hired by agronomists) responsible for analyzing and improving crop productivity.

“About 80 percent of the value we create is in the analytics,” CEO Michael Ritter recently told me. “We sell the hardware at a fixed price, with recurring revenue on the analytics.”

SlantRange raised $5 million less than three months ago from a consortium of investors in a Series A round of venture funding led by The Investor Group, a San Diego investment firm.

Ritter and co-founder Mike Milton raised much of their initial seed funding from farmers in Nebraska and Iowa. They set out in the fall of 2013 to test their system and collect data, and had planned to later seek venture funding in Silicon Valley. At first, the farmers thought they were crazy, Ritter recalled. Then they became investors.

“We’ve been doing this type of thing for a long time—airborne remote sensing,” he said.

Nearly all of the company’s 10 employees came out of General Atomics Aeronautical Systems, the San Diego-based defense contractor that created the unmanned Predator and Reaper aircraft for the U.S. military. According to Ritter, the SlantRange team has an advantage over most drone-related startups because they have a good grasp of the practical limitations of airborne surveillance from their work on sensors and other “mission systems” for the Predator.

“Data acquisition is extremely perishable,” said Ritter, who spent a decade at GA Aeronautical Systems. “A pest infestation can get out of control quite quickly and take over a crop. So making sure the data is accurate and immediately available makes it extremely valuable” to agronomists and farmers. Some rivals address this need for speed by designing their system to upload imaging data to the cloud. But SlantRange took a counter-intuitive approach.

The company developed a new technique for data compression that allows much of the data-crunching analytics to occur in SlantRange’s on-board sensor module and the tablet computer used to control the drone.

“Access to computing sources around the world do not line up well with agricultural lands,” Ritter explained, and bandwidth is a rare commodity. “The volume of data produced by these systems is just so large—maybe 200 to 500 megabytes per acre—that the data overwhelm the uplink capability” in most sparsely populated rural areas. “You need a [high-capacity] T1 connection to upload the data.”

In a field trial for a potential customer in the wheat-growing region of South Africa’s Northern Cape, Ritter said an Israeli-based competitor had to drive several hours to Johannesburg to transmit its data to Israel for analysis. It took several days to get the results to the customer, and SlantRange won the business, Ritter said.

In the meantime, Ritter said the FAA rules for operating radio-controlled drones by visual line-of-sight shouldn’t hinder SlantRange. Keeping a drone in sight usually means the operating range is “something just short of a mile for the types of systems we use.”

Ritter said American farms are usually comprised of multiple non-contiguous parcels spread over a large area. “Our typical Midwest customer is a 10,000 to 12,000-acre farm, which is comprised of perhaps 75 fields spread over several hundred square miles,” he said. Such jobs require the drone service provider to move from parcel to parcel anyway, so flying a drone within view usually isn’t a problem.

Ritter also maintains that low-altitude, slow-flying drones also can deliver very high-resolution images at a much lower cost than a farmer would get from manned aircraft or satellite imagery.

“By very high resolution, we mean resolution whereby the pixel size is smaller than individual leaves on the plant,” Ritter said. “Once you’re in that regime, you can begin invoking techniques in artificial intelligence and computer vision to deliver new types of information that are not possible from lower resolution manned aircraft or satellites. And that is where the real value is in multiple respects for delivering more specific and actionable information to growers.”

Bruce V. Bigelow was the editor of Xconomy San Diego from 2008 to 2018. Read more about his life and work here. Follow @bvbigelow

Trending on Xconomy