Autonomous cars are still a long way away from replacing human drivers on the road. But make no mistake: Intelligent, highly adaptable self-driving robots are here — and CANVAS Technology is leading the way.
CANVAS has been on our radar for a while. The Boulder-based robotics company launched in stealth mode more than two years ago, leaving us guessing at the work being done behind the scenes. It wasn’t until last September that the company finally revealed its mission: to build the world’s first completely autonomous, self-driving industrial cart.
Today, CANVAS unveiled its long-awaited brainchild and gave us a glimpse inside the complex and fascinating evolution of its technology.
CANVAS’ debut product will go to work in factories and warehouses around the country — long before autonomous cars hit the road. The carts are designed to automate the transportation of materials in industrial settings, using cutting-edge computer vision and real-time 3D mapping. CANVAS Autonomous Carts need no maps, require no infrastructure and function even in the fast-changing and unstructured environment of a warehouse floor.
“CANVAS Carts help companies enter the future of autonomous transport and stay competitive,” said Jonathan McQueen, CEO and co-founder of Canvas. “And the system can be up and running in less than 60 minutes, significantly improving processes without a large investment of time or manpower.”
But as flashy as the newly revealed carts are, the more impressive piece of the company’s story is the fact that all of the technology they use has been built from scratch.
When CANVAS launched in 2015, CEO Jonathan McQueen had a vision to make moving materials around a warehouse floor faster and more efficient. His co-founders were pioneering research and development in the realm of autonomous robotics.
The problem was that the technology they needed didn’t yet exist.
“It started with the cameras themselves,” Jonathan explained. “There were no cameras available with the performance characteristics needed to do this level of 3D mapping and navigation, so we had to create our own cameras.”
But it didn’t end there.
“The operating system that runs most robots — called ROS — was not powerful enough,” Jonathan continued. “So we had to create our own operating system. And then the most important piece was the map. You need a map that is updated by every robot, but yet those robots need to survive even when they’re out of touch from the central map. That map needs to be updated continually as things change but in a smart way — not simply updated with changes but with the implications of those changes. It’s quite a complex puzzle.”
From the hardware to the software, CANVAS has created a completely new type of robot capable of navigating intelligently while sharing real-time 3D maps with other robots in the fleet. Carts are able to learn from one another and adapt as the environment changes around them.
In 2015, when they sought the initial funding to bring the first autonomous cart to life, co-founder and VP of marketing Sandra McQueen said they were met with interested skepticism. Investors believed that self-driving vehicles, ones that could successfully navigate the chaotic environment of a busy warehouse, were five or even 10 years down the line — if they were possible at all.
Just one investor was willing to take the risk, and with that little bit of funding in hand, CANVAS set out to prove to everybody what they already knew: that autonomous robots could work, and they could work outside the structured confines of a test track.
“Robotics historically work in places with a lot of definition or structure, where it’s sort of predetermined what’s going to happen,” Jonathan said. “But robotics fail in the edge cases — the last 20 percent of places where unexpected things are happening. We’re actually focused on the unexpected cases. We like going into facilities where other systems would not be able to operate and finding solutions.”
The company’s carts are currently being piloted in industrial facilities that range from small distribution centers to massive warehouses and manufacturing plants. CANVAS will first focus on filling the backlog of orders from eager customers, and in the coming months, plans to ramp up manufacturing and distribution to meet increasing demands.