By Eric Peterson | Aug 07, 2017
Autonomous robotic carts
Industry: Industrial & Equipment
Products: Autonomous carts
While working at a composites manufacturer in the U.K., McQueen had firsthand experience with "this problem of moving materials through a facility," he says. "We humans put a lot of energy into moving stuff, particularly within factories and warehouses."
"The problem is moving stuff is a brainless task," continues McQueen, CANVAS Technology's CEO. "Nobody wants to do it. It's not fulfilling." Skilled people "want to move on from pushing carts or even driving forklifts."
Most industrial conveyance systems aren't nimble enough to keep up with manufacturing logistics, he adds. "They're big, expensive, and heavy items you can't move around."
Major manufacturers like Toyota have been using automated carts for decades, but they're not good with dynamic environments. "For half a century, these systems have improved very slowly," says McQueen.
The status quo for sensing, LiDAR (Light Detection And Ranging), sees in 2D. "They only see a fraction of the environment," says McQueen. About 15 companies manufacturer these kind of carts, and most of them run on the ROS operating system. "They're fine for a highly structured environment, but factories and warehouses are not static places."
Existing autonomous carts often have an involved rebooting process after they have an issue with changes -- say, a new pallet or a moved box -- in the physical environment. "Those systems get lost, and when they get lost, somebody has to reboot it," notes McQueen. "People on my team had no time for that. I dreamed of a flexible cart that I could load up in the warehouse and deliver to the line."
McQueen envisioned a user-friendly, smartphone-controlled, scalable network of autonomous robotic carts that were an order of magnitude more intelligent and adaptive than their predecessors, "something you could redirect in minutes instead of hours or days," he says. "More importantly, I wanted to move things physically and have the robots adapt."
That's especially critical in a factory or warehouse with constant movement and change. McQueen's solution, CANVAS Autonomous Carts, can carry 200 or 1,000 pounds; a cart with a 2,000-pound capacity is in development.
"They're practically invisible," says McQueen. "We don't get lost. You can change the factory and we know where we are. That's the difference."
To deal with disorder, CANVAS engineers developed "a new operating system, new cameras, and -- most importantly -- a novel mapping system” for the carts. Compared with 2D LiDAR, CANVAS maps in four dimensions, the fourth pertaining to change over time.
"To build that level of intelligence is really difficult," he explains. "It's been the Holy Grail of the robotics world for a long time."
For this reason, McQueen relocated to Boulder to collaborate with co-founder and CTO Nima Keivan, who'd come to Colorado so study computer science at CU, and has led the company's rapid R&D process.
Starting a company in Colorado "has been great," says McQueen, a California native. "It's just a supportive environment. It's probably like Silicon Valley was 30 years ago."
He says he was happy to set up shot someplace other than the Bay Area because of the constant workforce churn. "We'd be so far back from where we are," he says. Instead, the company has covered a lot of ground in two short years as launch approaches. "We've been running beta products for seven months now and refining and improving the carts."
McQueen says the beta sites are the facilities of "Fortune 50" companies, but he's looking at a broader market after the impending launch. "I came from a mid-sized manufacturer where it's really difficult to get an automated solution," he says. "My dream is to be able to supply the smaller manufacturer." To this end, a single cart will be "available for a low monthly fee," but larger manufacturers can easily scale up to have a relative army of autonomous carts.
McQueen is working with a contract manufacturer in Colorado and is looking for additional partners in other markets. "I really believe in manufacturing locally," he says. "I have a contract manufacturer lined up in Europe and we're in talks with a couple in Asia."
CANVAS Technology's startup has been financed by four California-based VCs: AME Cloud Ventures, Visionnaire Ventures, Morado Venture Partners, and Heroic Ventures. "We found the right ones who are knowledgeable and can help us," says McQueen.
McQueen says CANVAS Carts have the potential reshape and revolutionize workflow at facilities worldwide. "It's quite amazing to see how well these things work around people," he says. "I think it's disruptive. People are going to go, 'wow.'"
Challenges: "Our biggest challenge is hiring and building the right team," says McQueen. But he notes there's a cluster of robotics talent in Boulder and McQueen says Keivan helped attract top people.
"We have close to 30 engineers," says McQueen. "I think what’ most amazing is seeing how collaborative it is. I've never been of a team with no ego" -- and that selflessness accelerated R&D.
Opportunities: Growing into a massive market of manufacturers, e-commerce companies, and fulfillment centers. "We are growing rapidly and I think we'll continue to grow," says McQueen. The market is "really anywhere people are pushing carts. It's that simple."
He adds, "It's such a no-brainer. Instead of spending time walking, people spend time adding value."
Needs: Factories and warehouses in Colorado to test carts. "We would like more beta sites locally," says McQueen, noting that most test facilities are on the East and West coasts. "There are tons of warehouses and manufacturing in our backyard."
Meet CANVAS Technology at the 2017 Manufacturing Growth & Investor Conference, September 20, in Denver.