John Deere Interview: Robotics and Automation as an Enabler to Agricultural Systems Productivity
October 25 2011
by Mark Ingrebretsen, Editor, Robotics Trends
The past decade has seen enormous changes in the types of machinery used on modern farms. Drive-by- wire tractors, linked to GPS systems and receiving precise mapping data over the Internet have already led to significant productivity increases. But even greater increases in productivity will become necessary in the years ahead if farmers are to feed the estimated 9 billion people who will inhabit the planet by midcentury. John Deere is helping lead this push by exploring ways that robotics technology can make farming more productive and sustainable, as well as more rewarding for farmers themselves.
In this interview, John Reid, Director of Product Technology and Innovation, a leader of John Deere’s Global Technology Innovation Network, explains how John Deere began pursuing robotic technology and what changes that technology will bring to farming in the years ahead. A keynote speaker at Robotics Trends’ RoboBusiness Leadership Summit, taking place in Boston this Nov. 2-3, Reid came to John Deere in 2001 after a 14-year career at the University of Illinois, where he was recognized internationally for his contributions to robotic applications for off-road equipment. Reid’s expertise includes machine vision perception, controls, and hardware-in-the-loop design. He has received 20 patents and has published more than 186 papers, including 61 refereed journal articles.
RT: What is John Deere’s involvement in bringing robotics to agriculture?
Reid: We see that technologies that increase the productivity of our customers are going to be needed in the future. Up until now, we’ve been able to increase productivity by making machine systems wider and with more horsepower. This has been a good direction to get us to where we are today. But with the evolution of new technologies we see opportunities to follow a different path to higher productivity - one that involves automation-control and robotics.
Some of the fruits of that effort are already available. Since the U.S. government got the GPS network up and running in the ‘90s, our industry has been very heavily involved in using GPS positioning to automatically guide farm equipment. And this has proven a very valuable exercise at Deere, because it has helped produce the skill level required to accomplish farm-related tasks faster and more efficiently. The skill level also makes our products easier to use.
Our customers are more productive relying on GPS than they would if they were guiding our farm equipment entirely on their own.
RT: What are the chief advantages of GPS?
Reid: GPS is only one part of the picture. It gives you the accurate position. As a result, we’re able to use a sprayer applying fertilizer, and thanks to GPS we can drive more productively across that field, applying the fertilizer in precise paths, so there is less overlap and less under lap. Thanks to control and automation technology we can utilize information on what needs to be applied where.
For example, depending on topography or soil type, there’s a reason to have a different application rate. And our machines can control the application rate down to what’s released by an individual nozzle. Thus, the machine knows where it is because of GPS and it knows precisely where each nozzle is located on the machine as well. So applications can be made based on the site-specific needs in the field.
Before this technology became available, farmers would have simply applied fertilizers uniformly across a field which would not have been optimal for every specific location.
RT: Can you give me some examples of how John Deere’s machinery is using robotics technology right now and what plans you have to use it in the future?
Reid: Around the year 2000, the company began to see that there were some emerging technologies that when they matured, could have a dramatic effect on how we do business. A lot of those technologies are coming to fruition today. Improvements in sensors and telematics are examples. This led to the notion of seeing our machines as robotic systems. It really involves controlling our machines by automated, electrical signals. It’s called control-by-wire. Combining GPS with inertial systems has given us the ability to have precision guidance. That’s a development which has already proven successful in the marketplace.
Then we get to the harder tasks, such as planning the routes our machines take as they navigate over fields. About 10 years ago, Deere was doing research in this area, especially for patterned mowing of golf courses, stadiums, and lawns. Early versions of path planning technology is available today in something we call iTEC Pro. Using it, allows a machine to turn by itself at the end of rows.
In fact, there are certain things iTEC Procan do that human beings can’t. People tend to like to cover a field in adjacent paths. iTEC Pro can create a very efficient path, even skipping passes and then returning to them in subsequent passes to allow turning machines with reduced time losses. We’ve basically gotten to the point where you can just press a button and the machine will cover the entire field for some types of operations.
These are all innovations that have come about thanks to robotics technologies. The next level gets far more complicated, since it involves basically replacing the eyes and ears of the person operating the machine. John Deere’s first efforts in this area have been for the military. In collaboration with iRobot, we’ve developed something called the R-Gator. It is a robotic gator [ATV] that takes advantage of our very popular M-Gator military platform. By adding robotics capabilities to the M-Gator With this platform, we were able to demonstrate an ability to extend the soldier’s ability to operate the platform safely from a remote location.
Also, for about 10 years we’ve had a relationship with Carnegie Mellon University to develop perception technologies. These give us the ability to take a machine and know that not only can it be steered properly and navigate successfully over a variety of surfaces, but it can also see things that a human might normally see and take the appropriate response.
RT: Getting back to agriculture, what metrics do you use to measure the effectiveness of these technologies?
Reid: Our customers are seeing the benefits today with automatically-controlled systems performing more consistently and efficiently than would be possible otherwise. But up till now, they’ve been measuring productivity based on the machine’s productivity with the operator being a potential factor. Can this machine plant more hectares per hour if operated at a high level of operator efficiency?
Robotics adds flexibility to this equation. With the old paradigm, you could opt for bigger machines with more horsepower. And with automation, you could take out some of the variability between operators or due to operator performance. But with robotics you can create new machine systems that are optimized for the worksite requirements. With robotics, the new measure of productivity is the productivity of the worksite and what mix of machines and technologies you need to optimize it. John Deere’s FarmSightTM is an entry point to thinking of agricultural system productivity as a total worksite solution.
To illustrate a different type of value created by robotics, I’ll mention that Deere now has a robotic product on the market called the consumer autonomous mower, that’s just gone into production in Europe. Prior to robotics, you measured the productivity of your lawn mowing process by how long it took to mow it. I’ve had an autonomous mower at home in the past, and as a result, I measure success by the fact that my grass is maintained to look great even though I didn’t have to personally spend time mowing it. The autonomous mower (more of a “cut-length maintenance” system) itself works actually more slowly than a traditional push or tractor mower. But as long as it doesn’t take my time in maintaining my lawn to a high quality and superior appearance, the machine’s operating speed doesn’t concern me.
That’s a lesson which applies to how our customers make use of our robotic technology. If robotics adds a lot of unnecessary management complexity to their operation, then it’s not going to be as attractive as doing tasks other ways. But to the extent that we can automate those tasks in a way that results in higher productivity with less pain, then robotics technology will be accepted based on its ability to compete with the costs of solving the problems other ways.
RT: What do you see future farms looking like? Will they be vastly different as a result of your efforts today?
Reid: Agriculture always is undergoing change and the generations observe dramatic increases in productivity and convenience for the farmer. The automation technology changes today are dramatic and visible. Even 10 years ago, we didn’t have machines that were connected to the Internet or that possessed precision guidance systems. Today, we are at a point in time where there are numerous options and robotics is one of them that has a chance to prove out as a viable solution. But it has to be delivered in a way that increases productivity and reduces cost over other alternatives. The future is going to demand a much higher level of productivity out of farming systems. And they must lead to productive worksite solutions if we’re going to produce the food needed to feed the 9 billion people living on the planet in 2050.
