CES 2019 robots artificial intelligence

CES And Robotic Trends Of 2019: A Little Important Progress

Robby The Robot made its theatrical debut around the time I was born, and it set a bar regarding what a practical robot should look like and do. Lights and lighting effects aside, he could drive, lift heavy objects, cook, clean, and only seemed to have issues violating Asimov’s 3 Laws of Robotics and going up stairs. He was followed a decade later by the Lost in Space Robot which seemed far more limited in practical applications and with the overly creative name “Robot” (the new one is far more interesting). I’ve been a fan of robots ever since and the ring bearer at my wedding was a robot dressed in purple named Cedric.

At CES this year, we’ll be up to our armpits in robots from a variety of vendors—most falling short of those movie robots from decades ago. I think the issue is that we have lost the path to successful product creation by focusing too much on cost initially and not focusing first on setting the bar on what the market will accept and then cost reducing it in order to make it affordable. Instead the more common path appears to be to start with price and deliver something that falls short of expectations. I expect this will make the path to viable robotic assistants longer than it would be as this is the same process largely being used by VR makers now and sales haven’t been great in that segment.

Let’s talk about some of the robots that will define CES this year.

The Tēmi Personal Robot

The Tēmi Personal Robot is basically a tablet on wheels and it uses a design that was conceived some time ago to address one of the critical problems with video conferencing; side conversations. It has a reasonable price of around $1,500 and with its powered base it is moderately mobile (moderately means it can’t handle stairs or any objects a human could step over that would block its wheels).

The idea is that if you are remote and in a video conference you can move the robot around to have a private side conversation or something, we might today use texting for. While the concept did attempt to address a real problem, it never has taken off both because the remote user still feels remote and the person, they are talking to never seems to get over the fact they are chatting with a tablet on wheels. And unless you integrate headphones the side conversation is far from private as the device attracts attention like bacon attracts dogs.

A better approach to this problem is a product called Spatial which uses AR headsets like Microsoft HoloLens to place remote folks in the room more realistically. AR, particularly when we get to headsets that provide better occlusion, is likely a far better solution for this problem than a robot is. And while a tablet on wheels could be useful for the money, just carrying a tablet around is likely useful enough.

This is an excellent example of hitting a price point with a product that likely will have real trouble finding a market.

Cruzr from UBTECH

Cruzr from UBTECH is more interesting in that it actually looks like a robot with articulated arms. It even has hands, but it doesn’t look like they work. That seems like a very strange place to cut costs. The design reminds me a bit of the robot Marvin in Hitchhiker’s Guide To The Galaxy but, I hope, without the depression problem.

It is still basically a screen on wheels, though it is far more fascinating to look at and watch. It is significantly larger and likely significantly more expensive as well. It looks a bit like the security robots that have been around for a while and it does have a security feature set. As with the Tēmi, it seems to want to address that same video conferencing problem and it would have similar issues with doing it.

It has facial recognition but, with only 98 percent accuracy, it falls well below what Deep Learning systems can provide today (this could, however, be addressed with a 5G connection and a cloud service). Visually the product is impressive, and its AI capabilities do suggest it is farther up the robotic evolutionary chain than the Tēmi, but it too will likely fall short of what we want in a robot.

Intuition Robotics Elliq

The Intuition Robotics Elliq gets closer still to an ideal—mostly by shifting from something that is mobile to a more advanced Amazon Echo Show. Targeting an older demographic, it is designed to provide companionship both by anticipating the needs of its user and in more easily connecting the user to their family and care givers.

The tight targeting is what makes this interesting in that it doesn’t try to be more than it can be. For its target demographic it could be very useful, but here the lack of mobility could be a problem because people tend to move around their homes and unless you want one of these in every room (which is a viable solution) it would be out of earshot most of the time it was in use.

This AI and human emulation focus likely represents one of the big trends for personal robotics and I could see this actually meeting the needs for a single lonely elderly person in some cases.

Naver Labs Ambidex

The Naver Labs Ambidex is the closest to an ideal effort regarding utility. It has arms and fingers that can hold things, it is mobile, and it kind of looks like the evil robot in Disney’s Black hole went on a diet and lost he ability to fly. It does have a bit of an ominous look though and lacks the friendly appearance of either the Elliq or the Cruzr.

This unit depends heavily on the capability of the AI engine which is probably not yet where it needs to be, but this aspect, likely will be upgradeable. Particularly for someone that is disabled, this could be a lifesaver. However, one shortcoming is it is only part of the solution. It needs to be mounted on a mobile base in order to move around. If you added these arms to the Cruzr you’d have a complete solution.

While I expect this product is wicked expensive, it does have a viable purpose and it comes the closest to what I think a robot assistant should be. Particularly the care they’ve taken to develop safe arm technology. While this looks like it should be in a movie where the robots’ revolt with dire implications for humans, the fact is that the arms are uniquely safe and useful.

So, the most interesting solution isn’t really a solution at all but part of one which, I think, further showcases the problem.

Wrapping Up

There is a fundamental problem with the way we are approaching several markets, VR and Robots to name two. This problem is that we seem to not be focused on creating complete acceptable solutions but instead are too focused on price points resulting in products that aren’t attractive at any price. The failure of Baxter should have been a huge wake up call for the segment but, unfortunately, it wasn’t and that means we’ll again likely be up to our armpits in robots that few want to buy at CES this year. But the good news is, there is excellent progress and both the Elliq and Ambidex potentially showcase a path to far more successful future. The Elliq by focusing on human emulation behavior and the Ambidex by creating something truly useful for folks that need a spare set of hands.

Granted if I could just get something large to run around the house saying “Danger Will Robinson” that would scare the pets I’d likely be good.

Scroll to Top