NextGen Supply Chain: The robots are coming

It wasn't so long ago that people could generally agree on what an industrial robot looked like. They were typically big. Usually in cages to protect humans. And looked like a one-armed baseball slugger with a bat in his hand making the same motion swing after swing.

Some moved parts, others welded parts. Sure, there were smaller robots that did pick and place parts or palletized loads. But they were in the minority and not the first robot images to come to mind for most people.

Then along came Kiva (Gilt Groupe: No flash in the distribution pan) and Baxter. To begin, robots suddenly had names rather than model numbers. And they performed quite different functions in quite different environments.

Kiva bots raced around the warehouse floor moving shelving from storage locations to order pickers. No cages, no fixed position.

Baxter sits on the assembly or picking line next to humans and moves smaller parts. Again, no cage and much smaller than its ancestors.

Baxter and Kiva started to change the robot narrative. But it's now gone much further, just as general demand for robots increases at unprecedented rates.

North American sales of robots through the first nine months of 2017 (the most recent data available) were 14% ahead of 2016 in units, says Jeff Burnstein, president of A3, an automation advocate group. That growth was on top of 10% growth in 2016.

By all indications, we are just at the beginning of the transformation of robots. It's not just the number of robots but a rapidly changing profile of what they can do. Greater mobility and more smarts are the headliners here.

Over at Rochester Drug, a mobile piece-picking robot named Adam (first of his kind) roams aisles of inventory, selecting items directly from shelves without human intervention. Adam, from IAM Robotics, works at about the same pace as a human – 100 to 110 items picked an hour with greater accuracy. “The highlight of my night is when it misses something,” says Mike Collins, the human who monitors Adam and sends it picking instructions scanned from a bar code. Furthermore, Adam doesn't look anything like those big robots in cages. But it does need to have parts in the same place every time to complete the pick.

At Whirlpool, 54 robotic tuggers, each with a stereo camera-based navigation system from Seegrid, deliver washing machine parts to 24 locations on assembly and sub-assembly lines in Clyde, Ohio. “There is always labor savings associated with automation,” says DHL has tested collaborative mobile robots from Locus Robotics to pick parts at a warehouse in Tennessee. Retailer Hudson's Bay is doing much the same with RightHand Robotic's robots in Scarborough, Ontario. And the list goes on.

Locus Robotics Video

While many focus on the impact of robots on labor demands, smarts and mobility are the future. And while all of the applications just mentioned are significant breakthroughs, there is still another frontier being explored.

We'll start with smart. Traditionally, robots have required extensive, time-consuming coding to develop programs that guide movements not much more complicated than swinging a baseball bat. And by the way, unlike accomplished sluggers like David Ortiz, the robot's swing is always the same. There's just no adaptation to a changing environment.

That is rapidly changing. Organizations ranging from Elon Musk's OpenAI to MIT's Media Lab to the U.S. Defense Department's research arm Defense Advanced Research Projects Agency (DARPA) are working to advance AI-driven robots.

Pieter Abbeel, president and chief scientist at Embodied Intelligence, talks of using virtual reality to teach a robot how to move. Then artificial intelligence allows the robot to learn new moves on its own as the work environment changes. (See the NextGen Interview for more details .)

Then there's mobility, which is more than just a change in a traditional, repetitive robotic motion. Mobility is actually a matter of getting around the plant or warehouse, dispensing with the cage altogether. For instance, Jeff Christensen, vp of product at Seegrid, says his company's camera-based navigation system has logged more than 750,000 miles of travel on facility floors.

And that's only one type of navigation system for what are now known as autonomous mobile robots, which often look a lot like Kiva bots. Beyond stereo cameras, navigation technology here includes two-dimensional lasers, three-dimensional lasers and systems such as SLAM ( Simultaneous Localization and Mapping) and LiDAR (Light Detection and Ranging).

All of these technologies allow an autonomous mobile robot to receive an assignment, set its own path and not run into people or objects on the completion of its rounds. “If the navigation stack is not capable, then everything else falls apart,” says Simon Drexler, director of product development at Otto .

So, the future of robotics smarts is the ability of the machine to learn new motions to adapt to a changing environment. That's a long way from always making the same motion.
And the future of robotics mobility is being able to set its own path and navigate in a changing environment. That's a long way from remaining stationary either with or without a cage.

All of this has clearly caught the imagination of people in industry. Modern Materials Handling reported last year that 45% of survey respondents plan to upgrade or implement robotics in their facilities in the next 24 months. That's a substantial boost from the 20% that currently use robots in their facilities.

Or as McKinsey's analysts have said, “Today we are on the cusp of a new automation era: rapid advances in robotics, artificial intelligence, and machine learning are enabling machines to match or outperform humans in a range of work activities, including ones requiring cognitive capabilities.” Clearly, we're onto something big here.

Gary Forger is the special projects editor for Supply Chain Management Review. He can be reached at [email protected].

SC
MR