The Chinese company demonstrates its consumer bipedal humanoid at CES

This week at CES 2019, UBTECH Robotics (which was valued at $5 billion as of mid-2018) is announcing a major update to a walking robot first demonstrated at CES 2018. UBTECH’s Walker has gained a torso, arms, hands, and a head, and is now as humanoid as bipedal robots get. UBTECH has posted a couple of new videos, and answered some questions about Walker’s capabilities and where our expectations should be.

From the press release:
“Walker is your agile smart companion—an intelligent, bipedal humanoid robot that aims to one day be an indispensable part of your family. Standing 4.75 feet (1.45 m) tall and weighing 170 lbs (77 kg), the new version of Walker is more advanced than ever, including arms and hands with the ability to grasp and manipulate objects, a refined torso with improved self-balancing, smooth and stable walking in difficult environments, and multi-modal interaction including voice, vision, and touch. Walker has 36 high-performance actuators and a full range of sensing systems that work together to insure smooth and fast walking.”

There are a few things to keep in mind here. Walker seems to walk fairly well on a smooth surface free of obstacles, and doesn’t fall over, but that may or may not be a reasonable representation of how Walker would perform in an environment that is even slightly different. The piano playing is decent, but the fingers don’t appear to be actively actuated, at least in that version of the robot. I do like the box tracking and handoff, and it’s nice that Walker can sense and react to external forces, though I’m not sure we could say it is completely safe for human-robot interaction. It’s also worth pointing out that the push recovery, at this point, is very much best-case scenario: Walker is being pushed from the side while its near foot is planted; if pushed from the front or from behind while stationary, it may not fare so well.

I’m certainly excited to see what Walker becomes capable of as UBTECH continues its development, and it’s great that a consumer robotics company is investing in a humanoid like this. As we well know, building bipedal humanoids that can do useful things is very, very hard, and Walker is impressive even in its current state. Still, it’s important to be clear about what the capabilities and limitations of the robot actually are, especially when UBTECH suggests the robot “has the intelligence and capabilities to make a helpful impact in any home or business in the very near future,” and that it will “one day be an indispensable part of your family.”

For more on what we can expect from the robot in the near future, we spoke to Jeff Gordon, director of communications and PR at UBTECH Robotics, via email.

IEEE Spectrum: With recent failures of several social home robots, how will you effectively manage consumer expectations while describing your robot as “aims to one day be an indispensable part of your family?”

Setting consumer expectations is one of the most important things for any robot maker to think about, especially because the more human-like the robot appears, the more natural and intelligent people expect them to behave. Part of this is how we market the robots, but the greater challenge is the long-term advancement of robotics and AI. We believe we’re on a multi-year journey with humanoid robots, and as amazing as today’s robots are, the robots of tomorrow will be even more incredible and human-like.

What are the benefits of making robots more human-like?

We believe robots will, in the very near future (and in some cases today, like with Alpha Mini and JIMU Robot), play an integral part of our day-to-day lives—in our homes and with our families, at our places of work, and everywhere we go. When robots take a more human form, we interact with them more naturally like we would any other person; the human-like form eases their integration into our normal daily routines.

The press release says that UBTECH’s robots will allow humans “to focus our unique human ingenuity where it matters most.” Can you provide some examples of tasks that your robots are ideal for, along with tasks that humans will likely need to continue to perform?

Using UBTECH’s Cruzr service robot as an example, Cruzr is ideally engineered to be a first point of engagement when a consumer walks into a retail store. Cruzr can greet the customer, interact with her to determine her shopping needs, and even guide her to the location of inventory on the shelf. If, however, the customer has more complicated questions, Cruzr can activate video chat with a human call center, where a live agent can help provide answers, brainstorm solutions, or bring in an on-site agent for additional service.

Making bipedal robots that can reliably walk in semi-structured environments is very challenging. What will it take for you to be confident that Walker will be able to autonomously navigate around homes and businesses? Will it be able to recover from falls onto hard surfaces?

Walker features SLAM navigation and obstacle avoidance to determine the best path through a dynamic environment. It also has self-balancing capabilities so when Walker is disturbed by external impact or inertia, it can automatically adjust its center of gravity to maintain balance, or pick itself back up from a fall.

Manipulation in semi-structured environments is very challenging as well. What kinds of manipulation capabilities should we realistically be able to expect from Walker?

Walker is still in development, so more to come on this!

I appreciate that UBTECH says that “the long-term advancement of robotics and AI” is “the greater challenge,” and that they’re well aware that it’s going to take years. Walker is making an impressive debut at CES, but it’s got a long way to go, so we need to stay realistic about where it is right now. We did also ask how much Walker will cost, but it’s still in development so UBTECH has no availability and price information to share quite yet.