Robotics is one of the most exciting emerging technologies in the materials handling industry. Over three blogs, I’ll take a look at the development of a new mobile piece picking robot by a start-up with roots at Carnegie Mellon, and a pilot test at a pharmaceutical and medical supplies distribution facility. The columns aren’t meant to be an endorsement of any one company or its solution, but a look at one project.
Much has been written lately in the mainstream press about the future of robotics. Both the New York Times and the Wall Street Journal have noted how robots are growing in the workplace, as has the Boston Consulting Group.
The trend is also notable in the materials handling industry, where piece picking robots represent the next frontier. In Rochester, New York, startup firm Iam Robotics and Rochester Drug Cooperative are testing a new piece picking robot that, if commercialized, could be a harbinger of things to come. In this second of three blog, I’ll look at the development of the robot. You can read Part I here. Meanwhile, you can watch a YouTube video of the robot in action here.
Tom Galluzzo, CEO of the Pittsburgh-based start up Iam Robotics, didn’t go off to school to study robotics. “I was a space geek,” he says. “I started off as an aerospace engineer major as an undergraduate.” Somewhere along the line, however, he learned that robotics was a fast-advancing field. Sensing an opportunity, he switched majors and competed in events sponsored by the Association for Unmanned Vehicle Systems International. During grad school at the University of Florida, he was part of a team that was invited in 2004 to compete in the DARPA Robotics Challenge to develop a vehicle that could navigate its way across the Mojave Desert in Nevada.
“We had about two months from the time DARPA gave us the thumbs up to develop a basic GPS system for a 5,000 pound car that could drive in the desert,” Galluzzo recalls. “There were a lot of blind robots crashing into things that first year.” While one vehicle traveled 7 miles, most crashed before completing the first mile, including the Florida team. No vehicle completed the course. In 2005, five vehicles completed the 150 mile course, although the Florida team was not among them. “Basic GPS guidance and navigation was an issue our robot had back then,” Galluzzo says.
Fast forward and somewhere along the way, after a visit with Larry Doud at Rochester Drug Cooperative, Galluzzo got the idea for a piece picking robot. “We began talking with Larry a couple of years ago,” Galluzzo says. “We were coming from academia, so we really didn’t know what a distribution operation was like, but after our first visit, we knew this was something we could do.” The concept for Iam Robotics was conceived.
At the time of that visit, robotics had come a long way. For one, the problem of navigation had been solved, over and over again. “We had a lot of confidence in our ability to get robots to navigate in well-managed environments, like distribution centers,” Galluzzo says. The other important change was that the components, such as robotic arms and perception systems are much faster and much less expensive. The Xbox, for instance, has led to the development of affordable sensors and cameras that can see in 3D in real time. “What that means is that we have technology to get machines to see every day objects at human speeds,” Galluzzo says.
Humans are still far better than robots at recognizing and identifying the breadth of SKUs that might be found in many warehouses, but today’s technology can handle a set of 100,000 or more SKUs – say the slow-moving SKUs that might be found in a split-case pick module at RDC. As part of its solution, Iam Robotics has developed a high speed 3D scanner – “it’s like a CubiScan on steroids,” Galluzzo says – that can rapidly scan SKUs, create a 3D model, classify them and enter them into a database that allows them to be picked by robots.
In fact, the hard part is the ability of the robot to identify and grasp a part at full human speed. Galluzzo says he and his colleagues have solved that issue. “We’ve been able to get our robot to pick 1,100 pieces an hour when it’s positioned next to a shelf in the lab,” he says. “We believe the robot could do that on a sustained basis.”
While that kind of pick rate sounds impressive, the fact is, associates don’t generally stand in one place but move up and down an aisle in a pick zone. Galluzzo concedes that a person can move faster than a robot. “They will be physically slower than a person at a slow speed,” he says. The advantage is two fold. First, the robot can carry out the computational aspect of the job faster than a person being directed by a voice or RF solution. Second is that the robot doesn’t get tired – it can work at the same pace at the end of the shift as at the beginning of the shift – and it doesn’t get distracted.
“Our biggest hurdle is convincing the market of the robustness of a robotic solution,” Galluzzo says. “That is our burden to prove.”
There may be another hurdle, which is convincing end users to not only adopt robots, but to invest in a solution with a start-up. Many potential customers may recall that Kiva made a big splash before being purchased by Amazon and being pulled from the market. Galluzzo understands the concern. “What we’re doing to mitigate that concern is to hunker down for a long-term ramp up of the company,” Galluzzo says. “All of the founders are committed to staying with the company for a minimum of six years.”
So, is the adoption of robots just a matter of time? “Not in my opinion,” Galluzzo says. “It’s a matter of investment. The technology is here. We just need people to step up and adopt it.”