The rise of the robots
As robots get better, smarter and more adaptive, project teams have to keep up
BY STEVE HENDERSHOT ILLUSTRATION BY THE HEADS OF STATE
THE ROBOTS ARE ADVANCING.
Global robotics spending should grow from US$15 billion in 2010 to US$67 billion in 2025
Source: Boston Consulting Group
It’s not that they’re replacing humans. On the contrary, the latest generation of industrial and service robots is better able than its predecessors to interact with people. That improved compatibility, combined with greater adaptability and lower costs, has led to a proliferation of robotics projects: more product development, more installations to assist existing processes, and more midstream reprogramming and repurposing initiatives so that the flexible new robots remain useful throughout their life span.
In 2014, global robotics sales exceeded 200,000 units, the highest number ever, which surpassed the previous year’s largest-ever number of nearly 180,000 units, according to the International Federation of Robotics. The trend is expected to continue: Global robotics spending should grow from US$15 billion in 2010 to US$67 billion in 2025, the Boston Consulting Group reports.
“Breathtaking advancements and innovative technological developments”—such as faster, more reliable and more precise machines—have made robots easier to use and to integrate into industrial processes, says Gudrun Litzenberger, general secretary, the International Federation of Robotics, Frankfurt, Germany.
Even for experienced project practitioners who have worked with robots, the new breed presents fresh challenges and changes. “Advanced robots operate in more chaotic environments than traditional factory automation would put them in, including applications that require them to move inventory from place to place,” says Bryan Webb, COO and CFO of Kitchener, Canada-based Clear-path Robotics, which specializes in the design and manufacture of mobile robots for research and development, and industrial applications.
“Advanced robots operate in more chaotic environments than traditional factory automation would put them in.”
—Bryan Webb, Clearpath Robotics, Kitchener, Canada
The Power of Collaboration
The exemplar of the next-generation machine is the collaborative robot. Traditional robots, often sealed inside cages and isolated from the rest of the factory floor, performed tasks humans used to do. By contrast, collaborative robots interact with people rather than replacing them and might move from station to station, with sensors that keep them from running into anyone in their paths.
Robotics used to represent one big project—installation—followed by years of maintenance and then retirement. The new robots are programmable so that they can be endlessly repurposed, even to accommodate seasonal shifts in production—resulting in myriad projects that aim to reuse them.
A decade ago, a critical factor in any robotics project was the life span of the product that the robot would help create. The robot was built to do one task. Once that task was obsolete, so was the robot. That has changed.
“Often the training programs are left on the back burner with very low emphasis placed on actual handover to production.”
—Jack Mayer, PMP, Reis Robotics, Elgin, Illinois, USA
Consider the Versaball, a collaborative robot project completed in 2014 by Empire Robotics, based in Boston, Massachusetts, USA. Conventional two- or three-fingered robot grippers struggled to gently pick up objects of substantially different sizes and weights. The Versaball does the job with a balloon-shaped gripping mechanism that can carefully pick up objects ranging from heavy bricks to broken glass, then set those objects delicately inside a toolbox. One of the company’s customers uses the Versaball to put parts into a box at the end of an assembly line. It’s a classic robot task, but in this case the challenge is that the parts to be placed in the box change every week. Expensive, traditional solutions couldn’t adapt, but the reprogrammable Versaball can.
The Evolution of Innovation
Robots have revolutionized the worlds of manufacturing and beyond. By Patty Johnson
1961 The world’s first industrial robot is installed on General Motors’ production line in Trenton, New Jersey, USA. The 4,000-pound (1,814-kilogram) arm, named Unimate, follows step-by-step directions to sequence and stack hot segments of die-cast metal for use in the automotive-interiors plant.
1973 Japan’s Hitachi develops the automatic bolting robot for use in concrete pile and pole manufacturing plants. It’s the first robot equipped with dynamic vision sensors that allow it to identify bolts before tightening and loosening them.
1985 German engineers at KUKA develop the revolutionary Z-shaped robotic arm. The novel design achieves complete flexibility with its three translational and three rotational movements, which frees up space on factory floors.
1998 FlexPicker, the world’s fastest picking-up robot, is developed by ABB in Switzerland. Using image technology, the robot picks up 120 objects a minute at a speed of 10 meters per second (32.8 feet per second).
“The programming is getting easier, which makes the technology more accessible because there’s not as high a level of training required. It also makes the robots easier to repurpose,” says Bill Culley, president and co-founder, Empire Robotics.
“The programming is getting easier, which makes the technology more accessible because there’s not as high a level of training required. It also makes the robots easier to repurpose.”
—Bill Culley, Empire Robotics, Boston, Massachusetts, USA
The technical burden of overseeing robotics projects has lessened with the latest generation of machines. Robots are getting so good at understanding voice commands, for example, that soon people with no robotics experience will be able to program robots and integrate them into industrial processes, Ms. Litzenberger says.
While some technical hurdles may be decreasing, the project teams creating and installing these advanced robots still must accommodate the training of their end users. That’s been a challenge in robot installation projects, says Jack Mayer, PMP, a project manager at Reis Robotics, Elgin, Illinois, USA. “Often the training programs are left on the back burner with very low emphasis placed on actual handover to production,” he says. “You want a clear handoff so it’s not as if a robot just landed on the production floor, with people baffled by the strange thing that just showed up.”
2002 iRobot deploys its multi-mission robot, the PackBot, with U.S. troops for the first time. The robot analyzes suspicious objects, like potentially explosive material, allowing soldiers to remain at a safe distance.
2010 The first learning-controlled robot is launched by Japan’s Fanuc Robotics. The new technology allows the robotic arm to reach higher speeds and faster accelerations by intuitively memorizing vibration characteristics and suppressing motion, resulting in a reduction in cycle time.
2015 Pepper is the first robot that can read and respond to human emotions. Slated for release in 1,000 Nestlé stores across Japan by the end of the year, the humanoid robot will help sell Nescafé coffee machines by suggesting products based on customers’ responses.
THE FUTURE Researchers from the Chinese University of Hong Kong are building prototype microbots, each the size of a human cell, to deliver medicine into the body and eliminate the need for some surgical procedures. The miniature robots will be injected into a body without leaving wounds and will be controlled by electromagnetic fields.
Lower Costs, Greater Returns
Next-generation robots aren’t just more effective than their predecessors. They’re also cheaper. While robotic capabilities have climbed, costs have fallen at a pace “comparable to the way that costs of computers have gradually gone down,” says Bob Doyle, communications director of the Robotic Industries Association, Ann Arbor, Michigan, USA.
Volkswagen’s autonomous sales guide on display during the 2015 Detroit International Auto Show in Detroit, Michigan, USA
PHOTO BY ED ALDRIDGE/SHUTTERSTOCK
“Breathtaking advancements and innovative technological developments [have made robots easier to use and to integrate into industrial processes].”
—Gudrun Litzenberger, the International Federation of Robotics, Frankfurt, Germany
That means the return-on-investment analysis for a robotics project looks substantially different today than it did a few years ago. Costs are low enough that many companies are looking to break even within a year of their robotics investments. There’s more long-term value as well, because the increased adaptability and reprogrammability of modern robots means they’re less likely to be rendered suddenly obsolete as a result of a change to a factory process or a failed product.
Robots have been good fits for the automotive industry for years, but they are just starting to make similar inroads with other manufacturers—electronics makers, for example. “That’s because the same car rolls down the same line for five or six years with limited changes, but with electronics such as cellphones, the manufacturer changes the product much more frequently,” Mr. Doyle says. “Now, technology allows the robots to be changed to accommodate a different run application. That means better ROI.”
As a result, both unit sales and revenue are spiking. North American robotics orders totaled more than 21,000 units that generated US$1.2 billion in revenue during the first three quarters of 2014—up 35 percent in units and 22 percent in revenue from 2013, according to the Robotic Industries Association.
As new robots look different from their older counterparts, so too do the teams behind cutting-edge robotics projects. Current innovations in robotics are coming from smaller organizations such as Empire Robotics, and from partnerships between robotics firms and new customers looking to integrate robots into their processes. These new customers range from smaller manufacturers to giant corporations that operate outside the industries that traditionally have incorporated robots.
Here are two of the new-breed project teams that are pushing the boundaries of robotics.
Lowe’s began testing its retail robot at a Lowe’s-owned Orchard Supply Hardware store in San Jose, California, USA.
PHOTO COURTESY OF LOWES
At Your Service
A roving bot turns shopping in large stores into a pleasant experience.
A Lowe’s home-improvement store might look like a factory floor—with its concrete floors, fluorescent lights and roving forklifts—but the only assembly line here is the cash register. The 1,835 stores of the Mooresville, North Carolina, USA-based chain sell tools, appliances and hardware to 15 million customers each week, earning US$53.4 billion in fiscal year 2013.
These are not sites where one would expect to find a robot at work. Yet in 2014, Lowe’s completed a pilot project to create a robot that guides customers through its stores, helping them identify and locate merchandise within the cavernous spaces while speaking to customers in several languages.
To execute such a nontraditional project, the organization needed a nontraditional project team. In 2013, the company established Lowe’s Innovation Labs, an internal incubator that would generate outside-the-box ideas—far outside the box. Kyle Nel, the unit’s Mountain View, California, USA-based executive director, hired a team of science-fiction writers to imagine the future of the business. Then he had those ideas developed as comic books with characters and story lines, and presented the comic books to Lowe’s executives as project proposals.
“[A core strategy tenet is working with] uncommon partners—people and companies that you would never expect a company like Lowe’s to work with.”
—Kyle Nel, Lowe’s Innovation Labs, Mountain View, California, USA
That creative approach to identifying new possibilities is one of two core tenets of Mr. Nel’s strategy. The other is working with “uncommon partners—people and companies that you would never expect a company like Lowe’s to work with,” he says.
When Mr. Nel challenged his sci-fi team members to envision the future of the Lowe’s in-store experience, their ideas consistently included robotics. The timing was perfect, because Mr. Nel had just met an uncommon partner: renowned robotics engineer Marco Mascorro of Mountain View, California, USA. When the two men met in early 2014, Mr. Mascorro, co-founder and CEO of Fellow Robots, was working on robotics projects with retail applications.
Kyle Nel and Marco Mascorro, Lowe’s Innovation Lab, Mountain View, California, USA
An unlikely partnership soon formed between one of the 100 largest retailers in the United States and a new startup. Mr. Nel had met with several more established robotics firms, but in Fellow Robots he “saw this amazing team that was able to create unbelievable machines,” he says. “And I thought, ‘This is it. These are the people that are going to be able to work with us through all the muddy and hard parts to really make this happen quickly.’”
Still, Mr. Nel had to convince Lowe’s executives that it made sense to partner with a nascent startup on an ambitious robotics project. He did so by emphasizing his program’s mission of fostering innovation—a charter that implies a high appetite for risk. “I focused on the future, the vision, rather than all of the hurdles we had to overcome to get there,” he says. “Every time someone would bring up a potential hurdle, I’d say, ‘That’s true, but it’s worth the exercise and worth the risk if it can possibly get us to where we think we can go.’”
The partnership worked for Fellow Robots because it involved a high-profile project with a big-name partner; it worked for Lowe’s because the organization found a partner with the expertise to help it create new technology. “That marriage of an incredibly talented and driven tech company, along with a really large company, is a great match,” Mr. Nel says.
Mr. Nel wanted to get the robot into stores quickly, which meant committing to a short project timeline. Again, working with a startup proved advantageous. Mr. Mascorro’s team was willing and able to make the Lowe’s project an immediate priority.
Then came the development phase. The team leaders wanted the Lowe’s robot to serve some of the functions of a customer-service representative: to scan hardware, to process voice commands and requests, and to lead customers through the store to the locations of the products they sought. The primary challenge for Mr. Mascorro’s team was to combine an array of technological components into one coherent, useful, consumer-friendly robot.
PHOTOS BY TAI POWER SEEF
“I focused on the future, the vision, rather than all of the hurdles we had to overcome to get there.”
“We’re talking about speech recognition, autonomous navigation, computer vision [three-dimensional scanning] and a database,” he says. “We’re talking about putting all those things together in a way that has never been done before, and that’s been the hardest part.”
Consider voice recognition, for example. There’s no Siri equivalent for robots, Mr. Mascorro says, let alone a version that would result in a mobile robot wheeling itself around a store in response to a voice command. So the Fellow Robots team had to build one. It developed components such as speech recognition, scanning and locomotion, each in isolation, then worked to fuse those elements into one seamless robot.
In anticipation of unexpected roadblocks, Mr. Nel and Mr. Mascorro chose to treat their project roadmap as more of a loose itinerary than a fixed schedule.
“It wasn’t, ‘This is exactly what I need you to do, and you need to execute on this and make this happen.’ That’s a vendor-buyer kind of relationship, and in my experience that’s not how this kind of stuff gets done in the right way,” Mr. Nel says. Instead, Mr. Mascorro’s team regularly suggested project tweaks, such as adjusting the way the robot would scan store layouts.
The two team leaders stayed in close contact. “Open communication has made a big difference,” Mr. Mascorro says. “It’s kept the whole project moving faster, and we’ve been able to solve a lot of problems by brainstorming together. And because we spend a lot of time talking about the problems, when we came to points where we had to say, ‘We just can’t do that,’ they understood.”
As a result, the project advanced at breakneck speed. In November 2014, less than a year after the project’s launch, Lowe’s began testing its retail robot at a Lowe’s-owned Orchard Supply Hardware store in San Jose, California, USA.
PHOTO BY TAI POWER SEEF
“We’re talking about speech recognition, autonomous navigation and a database. We’re talking about putting all those things together in a way that has never been done before, and that’s been the hardest part.”
—Marco Mascorro, Fellow Robots, Mountain View, California, USA
For Mr. Nel and Mr. Mascorro, the prototype presented an opportunity to learn from its interactions with real customers and to learn about the constraints presented by real stores. For example, one of the surprising hurdles of the testing phase proved to be Internet access: The average Lowe’s store did not have a strong-enough network to accommodate the robot. Now the team knows the required network strength and has made changes so that a stronger wireless network is the only adjustment that stores will have to make to accommodate the robot. In addition, the initial prototype involved placing sensors throughout the store. Now all the required sensors reside within the robot itself.
The decision to introduce a robot to the public after less than a year of development, but in only one store, reflects the team’s agile approach: Iterate, test, refine and iterate again—as quickly as possible.
“We’re not approaching this project as having discrete parts, where now we’re done with version 1 and working on version 2,” Mr. Nel says. “We’re constantly going to be iterating and evolving and changing, and the course we have laid out is more directional than fixed. It’s going to change based on demand from customers, demand from store associates and the evolution of the technology. We’ll make advances, and then we’ll incorporate those into the robots as quickly as we can.”
Now, Mr. Nel says, Lowe’s is considering a rollout plan for other stores.
A Helping Hand
The Jeva robot assists people with physical disabilities.
For its very first product, 2Mar Robotics, a Melbourne, Australia-based company founded in 2013, set out to create a robotic arm marketed to an emerging robotics customer base: consumers.
Robots used in the home are expected to be the fastest-growing segment within the global robotics market, growing from US$1 billion in spending in 2010 to US$9 billion in 2025, according to the Boston Consulting Group. 2Mar’s robot, called Jeva, would serve as a household helper for people with disabilities.
Marita Cheng, the company’s founder and the project lead, began the development process with a learning tour. She interviewed people living with disabilities as well as organizations devoted to their welfare, asking for feedback to inform a prototype. 2Mar’s nine-engineer project team quickly assembled a prototype—a voice-operated, wheelchair-mounted bionic arm—and completed it in mid-2013.
The 2Mar team then tested the prototype in the homes of 15 people with upper-limb mobility issues. This first Jeva could lift only about 300 grams (0.7 pound). After seeing how the robot struggled to lift many of the users’ items, the 2Mar team decided to make the next iteration more powerful. Each machine costs more as a result, but its performance improved dramatically. The latest version of the Jeva robot can lift over 2 kilograms (more than 4 pounds). It will cost about AU$40,000, a mid-range price for robotic arms.
Ms. Cheng says her team’s process has changed and matured as Jeva’s development has advanced. Rather than moving as quickly as possible to completion and testing, her team now uses its insights from testing to inform a more protracted design phase for subsequent iterations. Spending more time upfront saves even more time, and costs, in the end.
PHOTO COURTESY OF 2MAR ROBOTICS
“In the early stages we just tried to get our prototypes done as quickly as possible,” she says. “Now we’re working on the fifth version, and we know a lot more about what works and what doesn’t work. The key is spending enough time in the design phase. The not-so-experienced designer may move onto execution more quickly, but they’ll find that their execution costs more, takes longer and doesn’t work.”
“We’re going to see much lower-cost robots, and we’re going to see them as much more valuable because they can do more.”
—Marita Cheng, 2Mar Robotics, Melbourne, Australia
Ms. Cheng’s team continued to test the Jeva robot up until the very end of the product-development phase in early 2015. That’s because she knows that the field of robotics is changing quickly and that in just a few years robotic arms are likely to become cheaper and more effective. Ms. Cheng is betting that, thanks to her team’s experiences and lessons learned while building Jeva, 2Mar will be prepared to capitalize on that less-expensive market as it emerges.
“What we’re building right now is still very expensive and still not accessible to the average consumer,” she says. “That’s going to change in the next five years. We’re going to see much lower-cost robots, and we’re going to see them as much more valuable because they can do more.”
PM NETWORK APRIL 2015 WWW.PMI.ORG
APRIL 2015 PM NETWORK