A small-time Iowan company has beaten the farm machinery big guns to the launch of a driverless grain-carting tractor that, with a bit of tinkering, could be adapted to cultivate, drill and spray without anyone in the seat.
Previously, such driver-free technologies have had a strange habit popping up for no more than a few months before mysteriously disappearing back into a research and development bunker, often never seen again.
Remember Case-IH’s cabless Magnum, John Deere’s autonomous grain cart or Fendt’s master-and-slave Varios? All have gone AWOL after a major early fanfare.
However, it looks like Smart Ag is here to stay.
On a modest budget, the tech company has built its self-driving system to plug the labour gap and offer farmers a way to keeping expensive kit moving when there just isn’t the qualified staff around to help.
Dubbed Smart HP, it sits above the tractor’s existing controls and allows external commands to guide the tractor autonomously.
Currently it’s only available for grain carting, with the tractor controlled from the combine by a web-enabled tablet computer that also works to highlight obstacles en-route and suggest precise paths to follow.
How does it work?
The AutoCart package includes four install areas.
The first is Smart HP, which cover the automation kit including all the wiring harnesses, safety systems, cameras and sensors along with associated hardware to connect the bits to the tractor.
Smart NX then connects your combine or any other machine to the cloud, allowing data to be stored and transferred.
Levels of automation
- Cruise control in a car
- Auto-steer in a tractor
- Adaptive cruise control in a car
- Supervised autonomous vehicle, such as the AutoCart
- Fully unmanned, such as a robotic mower
The software application – AutoCart – is the interface that the combine driver uses to control the field operations.
AAVI (Automated Ag Vehicle Intelligence) is the platform that deals with field boundary uploads, remote monitoring, data backup and account management.
All of these rely on seamless communication between the combine and grain cart.
The company can do this in three different ways – long-range radio signal, 4G mobile signal and wi-fi – depending on which can provide the strongest and fastest transfer of information.
There is also a separate antenna on the combine to send signals to the tractor’s GPS system.
Smart HP essentially works with data from three areas: telematics to detail where the machine is; the perception system around the tractor; and the vehicle’s interface for controlling the speed and brakes.
These three components work as a tripod. If one isn’t happy – for instance, if the network has dropped out or there is an unknown object in front of the tractor – then the machine can’t move until the issue is resolved.
In the cab
A harness is woven into the tractor’s wiring system, which allows the combine driver to take control once the Smart Ag button has been flicked.
Crucially, the tractor can still be driven as normal whenever the system isn’t engaged.
In work, the operator simply starts the tractor, leaves the gearbox in park and engine on tickover, flicks on the Smart Ag box and exits the cab.
The tractor is now fully controlled by the web device in the combine cab via the cloud-based Smart NX hardware using Google technology.
This orders the tractor to physically carry out commands such as steering, brakes and speed, which is limited to 15mph in autonomous mode.
All the software is developed and made in-house, with hardware sourced from local companies to save time and money.
There is no need to upload mapped fields to the software. Instead, paths are automatically created to avoid objects as the tractor is working from the field maps produced by the combine, so objects such as poles and ponds have (hopefully) already been negotiated around.
Unintentionally, the combine will always map the edge of the field first by taking off the headland and thus creates the external boundary, too.
The iPad will show where the machines are in the field and allows the user to draw paths between points and then press play to set the tractor moving and up the speed accordingly.
Alternatively, the driver can summon the tractor and let the SmartAg system automatically generate paths and work out its own speed – it has been taught to slow down for corners but work to its maximum on straight runs.
Cleverly, the tractor also knows that it can only go on areas that have been covered by the combine or another vehicle.
This eliminates the risk of the unmanned corn cart mowing down swathes of crop to take the most direct route to the combine.
There is also the option to set a middle point along the planned route to help avoid wet spots or such like.
The tractor has a designated home area and will return here once finished – typically it is at a gateway to unload into waiting bulkers.
Probably the biggest concern when selling a system like this is getting the tractor to function like a human, but without anyone in the cab.
A key feature was not to invalidate any of the safety features currently on the tractor.
That means that the hardware sits above the tractor’s Canbus system and all manufacturer warranties remain in play.
A human’s perception system allows us to see, hear, smell and feel what is going on with the tractor, whereas all the autonomous tractor can do is see/feel with its long-range, forward-facing proximity sensors.
These provide distance controls when the tractor detects an object, so at 10m the machine starts to slow and if it’s still there at 5m then the tractor stops and parks itself.
This was demonstrated by Smart Ag’s chief marketing officer, Justin Heath, who ran out in front of the moving tractor at a recent demonstration to prove that it would stop instantly at the first sign of danger and, crucially for the company, Mr Heath is still alive today.
It then sends an alert message to the user to say an object is in the way and won’t continue until the obstruction is removed and play is pressed on the iPad.
However, there is still work to do in getting the cameras and sensors to talk to one another in order to perceive the environment and not make such frequent emergency stops.
This will be vital in safely negotiating footpaths in the UK, for instance.
All workers are armed with emergency stops, along with one that is hardwired and placed next to the machine’s steps.
The ambitious company’s goal is to get the tractor to operating as safely as it would with a human driver, which relies on the sensors, cameras and technology working together without supervision.
Once this is possible, the use of the technology in other jobs such as cultivations, most likely as an additional subscription, will be possible.
The issue here is getting the tractor to think for itself with engine load requirements.
To do this, Smart NX is a crucial part and can be purchased independently from AutoCart to collect any type of data directly from the Canbus system.
This can be installed on any machine to harvest data such as yield, moisture, location and output, along with standard running data, and store it in a central location.
This type of software will be key in to increase the tractor’s autonomous functions in the future by using data to inform real-time decision making.
Currently, the Smart Ag system can only work with grain carting, where the combine operator is in control of the tractor’s movements.
There is also a need for someone to enter the cab in order to start unloading the chaser bin.
It is installed on a John Deere 8R (the job of fitting it takes about a day), but could be adapted to work on other tractor brands in the future.
There were 10 units working from Kansas to Canada during the last harvest and there’s a strong chance that the company will come to the UK but, for now, the focus remains on the North American market.
As for price, the fully installed system costs $35,000-$40,000 (£26,836-£30,670) per kit.
It’s not cheap, but could save the hassle of hiring substandard labour, especially in areas where the staff options are limited.
The small technology start-up based in Ames, Iowa, was founded in 2015 by Colin Hurd, who has been involved in agriculture and, more specifically, technology systems all his working life.
It currently employs 19 people, 75% of which are engineers.
Earlier this year, one of the US’ largest seed and grain growers, Stein Seed and Farms, invested $5m in the Smart Ag business, as it has been keen to solve the employment issue experienced by US farmers and saw driverless tech as an obvious way of doing so.
Watch Smart Ag’s video of the technology in action.