Mojow developing reactive camera-based system for self-driving tractors
An Edmonton-based autonomous agriculture start-up is developing a camera-based system that allows self-driving tractors to recognize and respond to changes in their surroundings.
Mojow Autonomous Solutions was incorporated in early 2020.
“We’ve developed an autonomous navigation and control kit that uses a camera for its main sensor,” explains founder and CEO Owen Kinch.
“It allows us to take an intake of 10 to 35 image-frames per second and analyze each and every pixel, looking for patterns, and ultimately what we’re doing is we’re creating a digital twin — a digital representation of the real world,” he explains.
As part of Mojow’s demonstration at Ag in Motion, a driver on an ATV drove in front of the self-driving tractor. The tractor recognized the obstacle and recalculated its path in real time to drive around the ATV.
The demo also featured the tractor transitioning a land roller from working field state to road transport mode and vice versa.
“We’re using the cameras as our feedback mechanism on how we control the steering, the propulsion, when to activate the remotes at the right time and step — very much like a human does, but our controller is doing it autonomously without the help of a human,” explains Kinch.
The controller is primarily communicating with the tractor via the CANBUS plug. “We take in that data, and we have our own sensors — primarily the camera, but also GPS and IMU, we analyze all that data in real time and basically get the job done in a changing work environment without any work stoppages.”
Mojow is developing a generic kit, but is also looking to build relationships with large manufacturers, he says. “We want to make a large impact, and we think through relationships with OEM partners, that’s going to be how we make the largest positive impact in ag,” says Kinch.