Exploring the future of farm robotics

By Tim Kalinowski, Staff Writer

 

With ag. tech. getting more sophisticated all the time, it is only a matter of time before autonomous robotic systems come to the farm. Already new tractors and other agricultural machines on the market include some of these features such GPS linking and navigation, precision seeding, spraying and header controls as well as ground contour technology. That next logical step toward truly autonomous machines is right around the corner. In fact, one scientist at the University of Saskatchewan has already gotten there.

Dr. Reza Fotouhi, an agricultural robotics specialist with the University of Saskatchewan’s Mechanical Engineering Department, says the reasons behind the drive for such machines is apparent.

“We need food. We have to also optimize food production. And robotics can help. The farming community is becoming a smaller group, and farms are getting bigger and bigger. In the future, maybe 10, 20 or 30 years from now, we will have a shortage of farmers, but we will always need to eat. So if we don’t have the manpower, we can adapt this technology to do this work for us.”

Fotouhi has built a working prototype for what such autonomous farm machines might look like in the future.

“The robot I have now is small quad about the size of a golf cart which can navigate and turn without an operator,” explains Fotouhi. “So the farmer can be in the control room and watch the robot navigating the farm.

“The robot is completely autonomous. It does not need a driver. We can give the robot a path through the field, or it can chose its own path. It can drive in the field, avoid the holes and it can stay on a straight line and seed back and forth. We are only focusing on the seeding right now, but it can be used for spraying and also for harvesting. These are just concepts. If in the future these systems become commercialized we could build a much bigger robot.”

Fotouhi engineered his machine using upper end, commercially available technology, alongside some innovations of his own. His goal was to create a model which could be available in the marketplace within five years.

“The machine that I have is not that complicated,” confirms Fotouhi. “It has four wheels, each with its own motor, that can rotate independent of each other. The left hand side wheels are linked together (by a program), and the same for the right hand side. That’s how the robot actually turns because it does not have a steering wheel.

“For navigation I put a GPS system on it. We have both a fixed GPS somewhere and a GPS on the robot so the two work together to get from point A to point B.”

To deal with obstacles in the field Fotouhi has built in a special 3D imaging laser system.

“I have put a fixed laser sensor on the front of the robot, and this is basically the eye of the machine,” says Fotouhi. “It does not have a camera. It only uses the laser to see, and it is extremely accurate because it can see up to 100 metres away. It also doesn’t depend on lighting conditions so it can actually work at night. The sensor can go up and down so it can generate a 3D image of what is in front of the robot. And then I map that into the computer.

“It can see bushes, holes or any obstructions in front of it with much greater accuracy than a standard camera system.”

Fotouhi also uses an industrial grade Internal Measurement Unit (IMU) to give the machine even greater navigational accuracy.

“I have an IMU because GPS tells you where your robot is; it doesn’t tell you which direction it is going or how fast it is going,” explains Fotouhi.

The GPS system, laser sensor and IMU combined allow the operator and the robot to chart its course precisely and safely through the field.

However, there are challenges facing greater commercial development of these autonomous farm machines. According to Fotouhi one of those challenges is cutbacks on grant funding for post-secondary institutions, which has essentially stalled further development on such robots in academia for the time being.

The second challenge is a general distrust of autonomous machines that have no driver physically present.

“It will take some time for the regulators and the legislators comfortable enough to allow these machines to go without the drivers,” states Fotouhi. “Other than this, if the grants are available to me, then I would guess in the next five years I could have the first generations of these machines come to market.”