Humanoid Robots Face Big Test With Amazon's Digit Pilots

Photo credit: Amazon

Amid a flurry of news, it was announced at this week’s Delivering the Future event in Seattle that Amazon will begin testing Agility’s Digit to bring the bipedal robot to its fulfillment centers nationwide. They are small steps, and such early-stage deals do not necessarily mean something bigger in the future.

Take Agility’s Ford pilot, for example, when the startup explored last-mile delivery as a possible path forward. Not too long after, the company began focusing Digit’s production solely on warehouse and factory work.

In April last year, Amazon named Agility one of the first five recipients of the company’s $1 billion Industrial Innovation Fund. While inclusion in the fund doesn’t guarantee that Amazon will use your technology in the future, it is a pretty clear indicator that the retail giant is at least interested in its potential.

“The Innovation Fund is really about exploring what’s possible out there,” Amazon Robotics chief technologist Tye Brady told me in an interview this week. “It’s also about understanding practical, real-world examples.”

The executive added that Amazon Robotics has only dealt in wheeled transportation, but legs offer numerous possibilities.

“We’re interested in walking robots,” says Brady. “I find it very interesting, the ability to move on different terrain is interesting. We’re also interested in what works about it – and, to be honest, what doesn’t. The humanoid shape is really interesting. I don’t know if it’s a good thing or a bad thing. We are experimenters at heart. We’ll find out. We’ll do a pilot and see how it works.”

The company’s focus on wheeled AMRs (autonomous mobile robots) dates back to its 2012 purchase of Kiva Systems, whose platforms formed the foundation for all of Amazon Robotics. There are currently 750,000 AMRs in use across the company’s warehouse network. The company has also launched non-AMR systems, including picking weapons like Sparrow, which was launched during the same event last year.

It’s hard to overstate the profound impact Amazon’s efforts have had on the rest of the industrial robotics industry. For one thing, the company has increased pressure on competitors to automate to meet growing customer expectations for same-day and next-day delivery. Second, the decision to stop supporting Kiva customers outside of the Amazon ecosystem directly led to the creation of some of the biggest names in the industry, including Locus Robotics and 6 River Systems.

For Amazon to integrate it into its growing robotics ecosystem, a system must demonstrate an increase in productivity. It’s less about innovation for the sake of innovation and more about finding possible advantages that result in goods getting to customers in less time. Including drones.

It remains to be seen how exactly humanoid robots in particular and bipedal robots in general might fit in. The other big hurdle is that any new system must accommodate the almost unthinkable size of the company.

There are currently a number of startups vying for the humanoid robotics crown, including 1X, Figure and Tesla. Digit from Agility looks the least human of all, but he has a lot of money and a huge lead. The company also recently opened a new factory in Salem, Oregon, which it says will be able to produce up to 100,000 Digits per year once it is fully online.

There’s no shortage of excitement in this category, but proving things at scale is a whole other ball game. Whether or not Digit succeeds at the tasks it is given could have profound implications for the trajectory of humanoid robots in general. Much like the Kiva systems proved to be a major catalyst for AMRs, if Amazon successfully rolls out Digit at scale, suddenly everyone will want to get their hands on some humanoid workers.

The biggest discussion point surrounding form factor is the fact that people build workspaces for other people. These include shelf heights, terrain, aisle width and the stairs, the bane of the ARM’s existence. From this standpoint, a humanoid robot suddenly makes a lot more sense. The reality is that most companies operate on brownfield sites. This means that their warehouses and factories are usually not designed for specific automation solutions. Humanoid robots fit well into an industrial wasteland.

Of course, Amazon has the resources to build any facility it wants, so it makes sense that many of its own robots operate effectively on greenfield sites. These limitations are less of a concern for Amazon than for most competitors, but if an effective system can be integrated into the existing workflow with minimal friction, that is of course ideal.

Photo credit: Amazon

However, Brady confirms that Digit is not the be-all and end-all of Amazon’s mobile manipulation plans.

“When you start bringing [sensing, compute and actuation] “Interesting combinations create really unique things,” he says. “We are the world leader when it comes to mobile robots. And now we’re intensively working on manipulating not only packages, but also objects. And it’s exciting to see all the possibilities when you bring them together.”

That could mean alternative paths. Amazon, for example, knows how to build both an AMR and a robotic arm. If you were to effectively mount the latter onto the former, you would have a kind of mobile manipulation on your hands.

“You see, with the agility robot you can think of it as a mobile manipulator,” says Brady. “We’re interested in that. The type of mobility is of particular interest to us as we have not worked much with bipedal robots. But definitely, we could combine this with identification systems, manipulation systems, sorting systems. We will do everything we can to innovate for our customers and improve safety for our employees. We’ll get there with the core fundamentals.”

If Digit fails to make the landing for some reason, it certainly won’t be the end of things, or for bipedal robots in general. Maybe it just doesn’t fit comfortably into Amazon’s existing workflows. Maybe the robot isn’t quite ready for Amazon size or Amazon isn’t quite ready for Digit to make sense yet.

Regardless, anyone even remotely interested in bipedal robots would be wise to pay attention here. The pilots could well have a profound impact on how we think about the category going forward.