KITCHENER, Ontario, Canada — An afternoon storm has dropped close to three inches of snow outside the Clearpath Robotics headquarters, but inside the industrial space, a flurry of beeps and whirs echoes every few seconds. The place is alive with engineering and code work, group discussions, and a handful of robots scurrying from one end of the floor to the other.
Welcome to one of the newer, and perhaps one of the more intriguing, spaces in the world of industrial robotics and automation.
Founded seven years ago during a robotics competition at the University of Waterloo, Clearpath has focused primarily on automating the “world’s dullest, dirtiest and deadliest jobs,” first in military and defense. That focus has changed, however, during the last year into advanced manufacturing robotics.
“Over the last 12 months, we’ve been working not only to hire the best minds that we can in self-driving vehicles and underlying control theory, but also to hire people with 15, 20 years of experience in industrial automation and manufacturing,” co-founder and CEO Matt Rendall said during a recent tour of the company’s facility. “Now, what we have is the ability to marry exceptional technology with deep domain expertise to deliver something specific and valuable to market.”
What Clearpath has is an automated guided vehicle (AGV) that acts a little like a robot and a lot like an autonomous car. The OTTO 1500 can haul as much as 1,500 kilograms (hence its numerical name, a weight equal to 3,307 pounds) up to 4.5 mph across factory floors, designed for continuous operation. Its initial deployment is focused on assembly line operations, and assembly line participation. Though still in its early stages, it has the potential to overhaul material handling and open the door to automation on the floor to even more companies. Some of the early users and investors include John Deere and GE.
Rendall fielded some questions about the company, its technology and its future.
IndustryWeek: Clearpath Robotics has been around for seven years but you only recently shifted your focus to advanced manufacturing. Why switch now?
Matt Rendall: We were building robots in university and immediately caught the robot bug. We became very passionate about taking self-driving vehicles into a business. At the time, 2008, 2009, most of the investment happening in autonomous vehicles was taking place in ... defense and aerospace, so we started offering development platforms to help accelerate R&D in military, mining, agriculture, and we really got exposed to every use case in every industry for self-driving vehicles.
The reason why we focused on manufacturing and logistics, quite simply, was because the market was pulling us the hardest there. There are a lot of interesting opportunities there in flexible automation in a manufacturing environment, and the customers weren’t seeing off-the-shelf solutions available, and they were coming to us for our research. We started studying the problem, trying to quantify the market, the opportunity, and the competitive landscape, and eventually, we made the decision to focus the company on material transport using self-driving vehicles. The short answer: Customers told us they needed it.
IW: What has driven the transition and your current OTTO projects during the last year?
MR: One of the biggest challenges we had was, we always knew we were good at self-driving vehicles, but it’s marrying the self-driving vehicle with deep domain expertise to solve an actual problem. … If you look at material transport, what are the most important things to understand? How much can I carry? And how fast can I take it where I need to take it? After that, it’s, Well, my vehicle needs to be able to respond to changes in the environment. If a forklift comes out of the aisle, the robot needs to be able to move around the forklift and continue on its way.
There’s been a big trend in robotics over the last three to five years: collaborative robotics — how do you get a robot to interact safely with a human in an unstructured environment? — and we’ve taken it to the next step. OTTO can freely roam in a factory and maintain industrial safety while collaborating in direct proximity with humans. I think the evolution of collaborative robots is self-driving vehicles.
IW: You mention Google and its autonomous car program often during presentations and conversations as both inspiration and, seemingly, a bit of a measuring stick. I would think that at least the basics of the technology used by Google for that project and the technology used by Clearpath are relatively similar.
MR: If you look at the Google Self-Driving Car, you’ll notice a big, spinning laser on top. That laser is capturing a three-dimensional view of the world. That’s the primary sensor, and there are a couple other sensors that will help increase awareness of the surroundings and help deliver the true self-driving operation. The fundamentals of the technology is something called SLAM — simultaneous localization and mapping. Basically, we figure out where we are while mapping the environment. Using the sensors we have onboard OTTO — primarily, a two-dimensional laser on the front and the rear — we construct a map of the environment and, at the same time, figure out where we are on that map.
One way to think about it is we spoof GPS. There’s no GPS indoors, but we can create our own GPS, so our control system is able to say I want to go to the x and y coordinate over here because there’s work to be done. The system is able to find its way to the x and y based on the spoof GPS. Using the same sensors, not only are you able to emulate the GPS system, you’re able to detect when there’s an obstacle and move your way around it. We can adapt the speed of the vehicle based on whether something is in front of it and have it move around the object. SLAM allows for that.
IW: All right. Important question: If I want to work with you to implement a fleet of OTTOs on my floor, how much?
MR: I can tell how it works, but I can’t tell you how much it costs to produce.
IW: Who have you been bringing in for this transition to the manufacturing realm? More roboticists? Automotive veterans? Folks from other industries?
MR: We’ve been hiring out of the automotive supply chain and the automotive industry. One of the really exceptional things about southwestern Ontario is the automotive corridor, and the automotive corridor has suffered over the last five years. There is this exceptional pool of talent with a tremendous amount of expertise in an area that is very valuable to us. How do we pull that capability into Clearpath and deliver new advanced manufacturing capability into that industry? Industrial automation, automotive supply chain and manufacturing are three areas where we’re focusing on building that deep domain expertise.
IW: Have you brought in folks from outside Ontario, too?
MR: One of the biggest challenges we’ve had is how to convince the world’s best robotics talent to come to Kitchener and Waterloo in Ontario, Canada. Historically, we’ve been very successful in recruiting people from around the world — Canada actually has a much more accommodating immigration policy, so we’re hiring PhDs out of India to come work with us. The biggest challenge for us is how to get the best talent out of the U.S. to come work with us in Canada. If we’re talking with people who are closer to the border, it’s a little easier, but what about the Bay Area? Our response was actually to open a Bay Area office in Sunnyvale, (California), and now we have a presence in the Bay Area, while we maintain our engineering headquarters here. We’re adapting to where the talent requires.
IW: Beyond the perks of working for a growing company (Clearpath anticipates growth this year from 125 employees to about 200) and interesting projects, how do you sell working here?
CTO Ryan Gariepy: When it comes to larger companies that may be doing some things that are more experimental, in many cases, robotics may not be that company’s reason for being. If there’s robotics talent that’s interested in doing this specific work, we can guarantee, if you come here, you’re working on robots. If your project is shut down, if this project is shut down, it’s because all of us lost our jobs, too. You can come to us and see there’s clear company commitment to really pushing the state of the art. With larger companies, that experiment may not exist in two years.
MR: We built the core robotics team a long time ago. Now it’s building out the pillars around it, because for a robotics-centric company to be successful, there are a lot of other areas that need to be in line. We’ve already hit critical mass, we already have all of those functions, and it allows us to take a complete robotics product into market in a way that I think will take other companies entering into this space a little while longer to build up, especially if they’re only hiring roboticists.
