Anthropomorphic Robots

I have to admit to a certain amount of skepticism when it comes to robots designed to mimic human modes of action. For years, I've had to explain to friends outside the industry that anthropomorphic robots were almost never the optimal solution for industrial automation.

So, here's a company with a product that claims to have created a robot that can co-exist with humans in a workplace, side-by-side, without guarding, light curtains, etc.

This isn't exactly anthropomorphic -- in fact, its arms look fairly conventional -- but it's alleged to be trainable by "demonstration", meaning you physically move its arm and hand to teach a motion.

Now, I seem to recall other robots being introduced previously that are similarly trainable, and I have to imagine that real automation problems are a *lot* more complex than can be addressed in such a rudimentary manner.

The main distinction seems to be that they've added extensive sensing to detect imminent collisions with any nearby humans. Note also that the robot features a display in the location of a human head, with cartoon eyes -- just too cute.

Does anyone have any experience with these, or similar, systems? Does any of this sound plausible?

Ken Crater
I just saw this robot in an advertisement email from one of our vendors.

For the work we do it would be a laughable to think that this would be productive. I've seen the gross negligence that very technical people have done. If we get to the day that non-technical people are training robots I don't think I want to be working for automation anymore.

Ken E said...
> For the work we do it would be a laughable to think that this would be productive.

Yes, that was my impression -- it seems like it would mainly be suitable for simple pick-and-place applications, for which it would be overpriced. I saw a mention of a vision system, but I can't see how that would be programmed by simply moving the arm.

And, I have to say, the cartoon eyes are just creepy.

Still, I can see a non-technical manager reading this and thinking, "I could replace workers for just $4 per hour", and assuming a bunch of magic that just doesn't exist.

Did this just make life harder for everyone trying to sell real solutions?
I think this company is targeting non-technical and "trendy" managers. There is a reason industrial robots have cameras often mounted somewhere on the arm with special purpose lenses and complex lighting. Often the lighting is behind the workpiece so as to put the part in profile for max feature contrast. The real applications tend not to benefit from humanoid appearances, but are determined by function. Maybe there is a niche market for this and I shouldn't come down so hard on them. I do have concerns whenever something is marketed as being simplified for the masses that it puts a negative spin on our industry when the rubber meets the road.

About 3 years ago I received a phone call from a small business owner complaining that the PLC he had purchased from us was not working. After asking all of the usual questions and realizing that this guy was not "technical." I asked who programmed the PLC for him and could I talk to that person? The answer "What program?" said it all. This person thought that if he wired the PLC to his process it would "just do it".

An example of marketing hype going too far.

Looks like this Robot has a similar path to follow.