Make use of the massive data from the seeder, sprayer and combine as the internet of things digs into the dirt
The third heaviest component in modern cars is the wiring harness. One-third the cost of building a modern car is assembling that harness. The task cannot be handled by robots.
To compound the problem, wiring harnesses are becoming more complex as they handle more information from more sensors and sources. There are accident prevention sensors along with sensors to manage the engine and drive train. They’re nearly all inter-related and they’re nearly all beyond the capability of robots to assemble.
Manufacturers of cars, trucks, heavy equipment and farm implements all share the same harness hurdle.
Many companies contract the work out to developing economies where labour is cheaper. Rigid testing protocols determine whether or not each harness has been correctly built. Quality control is vital when assembling a harness. One simple mistake can turn into a fatal mistake.
The same type of digital wiring labyrinths are integrated into modern seeding equipment, sprayers and combines, says Scott Day, Director of Agronomy at Fall Line Capital, a private equity firm focused on investing in ag tech startups. Day is also the fifth-generation operator of the family farm at Deloraine, Man.
Day says Fall Line recently made an investment in a California startup called Ethernovia, a company dedicated to exploring new technology for managing the large volume of information now generated by modern vehicles and equipment. Day commented on his impressions of the work carried out by Ethernovia.
“It’s really interesting when applied to agriculture. You’ll be able to process massive amounts of data on the fly. Our current optical imaging systems are now at their absolute maximum in terms of handling data,” said Day.
“If Ethernovia works the way they’ve described, handling massive data volumes will become reality. It will become normal to spray individual weeds in a field, because your sprayer will process optical images so quickly. The same instantaneous response can be brought into combines and seeding equipment.
“To me, it’s almost seems like they’re creating a new protocol for the way sensors communicate through the ethernet. For the immediate future, Ethernovia is looking at the automobile industry and the space industry. Fall Line saw the potential of Ethernovia in agriculture and that’s why we invested. We hope to be a resource to them as they explore ways to apply their technology to farm equipment.”
As one combine specialist commented recently, “If farmers today think their machinery is complex, just wait until the next generation of combines come along.”
Ethernovia expects to be there when those combines do come along. Ethernovia was launched in 2018 by Ramin Shirani. His experience in developing ethernet technology goes back to 1991, when he invented and patented the auto-negotiation technology, which is now employed in billions of ethernet devices worldwide. Auto-negotiation is the initial ethernet handshake between two computers.
“Farming is an important aspect of our market. However, right now the bigger market will be personal automobiles,” Shirani says in a recent phone interview from his office in San Jose, California.
“All the new cars, trucks and implements are being driven by next generation sensor technology. You have cameras, you have high resolution lidars, radars, ultrasonic other sensors. They gather data for perception.
“Sensors observe the field for weeds, soil fertility, moisture and all that. Cameras are having higher resolution, which makes their imagery more valuable. The data output is going significantly higher. Within the machine, there is movement of data from all these different sensors.”
Shirani says data flows to centres where it’s processed for artificial intelligence or for recognition of what the camera sees. Every OEM has their own proprietary architecture. There’s no standardization in how the data is collected, processed and turned into a physical act. The same challenge faces automobiles, trucks, airplanes and farm equipment.
“For example a BMW coming off the production line in 2020 has a significant number of sensors on separate networks. It might be traditional 30-year-old CAN or any one of a half dozen different networks, including some ethernet. There might be as many as 86 distinct networks that do not talk to each other and as many as 30 CAN networks and they won’t talk to each other either. It’s a mish-mash.”
He says the labour required to work on wiring harnesses is a major bottle neck that must be resolved.
“The cleverness comes from how we consolidate networks, while respecting the legacy networks that are part of all vehicles. We already know they work and they’re cheap. To accomplish this we use ethernet, which is now 25 gig and 50 gig and up to 100 gig. The car becomes a data centre on wheels.”
But automobile engineers don’t want hundreds of pounds of wire to pull these networks together, nor do ag equipment engineers, says Shirani.
“To make this work, we want to single pair two wires, a twisted pair. We want to be able to receive and transmit data at the same time, because that’s how duplex ethernet works. In a data centre, weight wasn’t as critical so the number of wires weren’t as critical. So a 10 gig ethernet in data centre runs on four wires.
“But in a vehicle, you can run 10 gig over a single pair. So that same wire has to transfer data at 10 gig in one direction and 10 gig in the opposite direction, at the same time and on the same wire. That’s full duplex. That’s how you eliminate redundant wire and at the same time transfer data faster.”
“Nothing Ethernovia does is proprietary. The minute you have something proprietary, people in the industry walk away from it. Users don’t want to be stuck with a single source. So everything we do is standard-based.
“We’ve talked to John Deere. I can’t say much because we have a (non-disclosure deal) with them. They use a lot of CAN, which is very common in agriculture and trucks. But it has limitations.
“With this NDA, I can tell you we’re familiar with their booms that have 48 high definition cameras. How you gather all that data from so many cameras and how you consolidate it, how you send it for logging, and how you process the data, is all a challenge. All the while camera resolution is going up. My understanding is they’re trying 96 cameras on the next boom.”
Shirani says running ethernet on single pair wires is the backbone that will allow the gathering and processing of the massive volume of data coming our way.
“In the next six to 12 months, we’ll sample our technology to key players in agriculture. Once we can demo the integrated circuits, duplex ethernet proliferation will move into the agricultural sector.”