Build Trust in Tech Before Autonomy

Trimble Working Toward Trust in Fully Autonomous Farming Technology

Dan Miller
By  Dan Miller , Progressive Farmer Senior Editor
Trimble is developing applications that will feed ever-greater levels of autonomy into farming. (Photo courtesy of Trimble)

Autonomy is not an end goal or even a specific point in time. It is a result of small technological steps, an evolution over years where trust in technical systems reveal new efficiencies and management systems. Autonomous planting, spraying and harvesting by way of a remote, hand-held tablet, for example, are not here yet. But they are coming. We recently spoke with Trimble about that day and how agriculture may get there.

Trimble's development work is based on four levels of autonomy. They are:

1. Operator assistance: The operator is reliant on visual indicators to provide information in real-time, for example, collision avoidance.

2. Task automation: A smart system capable of automatically completing tasks (establishing lanes used by a variety of vehicles).

3. Supervised autonomy: The operator is in a supervisory role but is still responsible for reacting to unexpected conditions. This starts with an operator in the cab in a hands-off, feet-off environment, moving into level 3 (eyes off).

4. Full workflow automation: Full workflow automation includes intelligent autonomy, optimized for a business need and operating without direct supervision.

While Trimble is working across the autonomous spectrum, the company is focused on solutions that will move farm managers from level 2 to level 3 -- where operators function as supervisors but who are still present to manage unexpected conditions.

DTN/Progressive Farmer spoke recently with two Trimble autonomy experts, Jeremy Leach, director of autonomy for Trimble Agriculture, and Kevin Andrews, strategy manager for autonomy, Trimble.

**

DTN/Progressive Farmer: Give us a realistic view of the space autonomy occupies today in agriculture?

Leach: The levels of autonomy for ag are real, and they're all significant. But they can also seem very, very small at the time. Guidance is one of those steps. We've become so used to it that if we took it away, people would be upset because it's invaluable. A level 2 solution offers assistance to the operator. In order to get to a level 4 solution, you need the level 2 solutions to be in place, you need the operators to have confidence in them, so that they trust in them.

DTN/Progressive Farmer: Would you give us an example of a level 2 solution?

Leach: Level 2 is hands-off, feet off. You don't necessarily need to steer the wheel. And you don't necessarily need to control the pedals. But you do need to be there. Level 3, we call it "eyes off." You don't have to be watching everything all the time. As an autonomous solution, it is actually able to provide you a higher order of assistance. But you can't completely abandon it to its own devices. An eyes-off scenario means you may still be in the field or at the side of the field, watching what's going on, or potentially going on behind the vehicle to make sure that it's doing the job that you would expect it to do. Level 4 is attention off. The work can be done without the farmer being present, meaning it could be done at night while the farmer is having lunch or attending to other critical work. Level 4 includes an understanding of the work that needs to be completed and the ability to complete the work under both normal operating conditions and in the event of reasonable exceptions. When an exception is encountered, level 4 machines have the ability to adjust so that they can complete the job.

DTN/Progressive Farmer: Agriculture is moving from a level 2 system of automation to a more autonomous level 3 system?

Andrews: The concept of levels was written for the automotive -- you are trying to get from point A to point B and using a car to do it. But remember that ag customers and farmers and operators have a much more complicated job. To be able to make that jump from purely operator assistance to purely eyes off is massive. We're adding a little bit of intelligence to the implements, and we're adding a little intelligence to the power unit or tractor, we're adding a little intelligence with our management system, and each of these things are taking gradual steps towards greater levels of autonomy.

DTN/Progressive Farmer: Is there a risk that farmers get overwhelmed with the technology?

Leach: It's not about just throwing on more buttons and bells. We've got to make sure that the user experience and the farmer's experience are optimized, efficient, and it's something that they're comfortable with.

Andrews: If it makes farmers' jobs harder, they won't use it. That is very much in mind as we're on this journey -- that it works as promised and does that every single day, because if it breaks down and they don't have a plan B because we've replaced everything with automation, then this isn't going to work.

DTN/Progressive Farmer: How would you describe Trimble's strategy toward autonomy?

Andrews: The Trimble strategy hasn't changed much in that we are building and absolutely investing in autonomy. Trimble has released something with IHI Corp that's outside of agriculture but will very much find its way into agriculture. IHI manufactures things like aircraft engines, and they have to move material from one facility to the next, and they want to do that very, very precisely. It's exactly the same way that we keep a tractor on a precise line. But their environment is very difficult. They've got tall buildings, they've got industrial facilities, they need to go into a warehouse. They need inertial navigation. But even inertial navigation isn't perfect. So, we've layered on laser scanning as a positioning method. LiDAR improves the positioning so we can keep that precise guidance going in and out of buildings. Imagine that you need to navigate under a tree canopy or you're in a dense forest or something like that. This is all highly relevant to ag.

DTN/Progressive Farmer: Trimble's work with IHI is really about connectivity, or the lack of it among buildings and inside buildings in an environment that doesn't lend itself to connection.

Leach: Connectivity is critical, but it does not have to be continuous. And let me explain that. We are very aware that not all fields in America, let alone in the world, have cell coverage or high-speed internet coverage. We know that our solutions need to be able to run in an environment where they're not connected. This means that they need the ability to pull the necessary information down to the vehicle in the form of a job (work order) detailing what is supposed to be done. As a result, they have the ability to operate in the field without connectivity. When something occurs that is outside the realm of expectation, the action taken will depend on the level of autonomy that is supported -- at low levels, the machine will rely on the human to intervene and make a decision; at higher levels, the machine will be able to make adjustments in order to finish the job. And then when they finish that job, they need the ability to upload the work that was done. Connectivity is critically important, but it doesn't have to be continuous. You don't have to do that in real-time; it can be done back at the farm rather than in the field.

Dan Miller can be reached at dan.miller@dtn.com

Follow him on Twitter @DMillerPF

Dan Miller