In my post on what makes a robot look feminine, a friend of mine posed an interesting observation in the comments. He felt too much time was being spent on humanoid development. He’d rather have development focus on non-humanoids that are useful now, at his home, than watching prototype humanoid robots through a glass window at the lab (not his words, I’m adding a bit of color). While this is a very understandable opinion, it’s based on a false assumption: Time spent on humanoids is miniscule in comparison to the entire range of robotics.

The domination of humanoids in robotics is illusive. Humanoid robots are a minority of robots under active development. Without any references, my guess is that they account for under 10% of all robotic development (if that). But in my experience, its definitely not uncommon that people perceive the exact opposite.

The reason for this is manifold. First and foremost, the media is very eager to write stories about humanoids. Humanoid robots have a lot more features that people can discuss and relate to. (Terminator references being horribly frequent). By following mainstream media or popular science stories, its quite logical to assume that humanoids are dominating robot development — while in fact, they’re just hogging all the attention.

People love hearing about robots that look and act human. Whether you’re a Terminator fanatic driven by a sense of self-preservation in a frantic opposition of robotics, or if you’re a robofan like myself who considers artificial intelligence the future of mankind — everybody’s interested in humanoid robots. To elaborate on my earlier point, robots with humanoid features make it a lot easier for people to perceive them as intelligent: Head, eyes, arms, legs — these imply “living being” whereas a rectangle chunk of metal on belts implies “vacuum cleaner”.

Humanoid form also allows people to ask questions right off the bat. Because the robot has the same form as we do it’s consequently a lot easier to imagine what it’s supposed to be able to do: Can it fetch my beer? Can it wash my dishes? Can it jump? Can it play soccer? Can it answer all these questions by itself?

This increased enthusiasm of the public both incites journalists to write articles on them, and, this is also why a lot of toymakers are trying to produce viable humanoids for the entertainment market: They’re very easy to market because people instantly “get” what the product is and does.

Another reason for why they get so much media attention is that humanoid robots pose a lot of difficult problems: balance, hand-eye coordination, object manipulation, etc. So, most humanoid projects are dealing with a very broad set of intellectual capabilities. These kinds of broad and intelligent systems are the cream of the crop, and consequently the media is eager to talk about them in a flashy manner. For example, if a project is dealing with humanoids manipulating fragile objects (a very multifaceted problem) — we could expect headlines along the lines of “Scientists getting closer to a robot that washes our dishes“.

To sum up, humanoid robots are in a minority. Media attention is naturally biased — but justified in the sense that humanoid robot projects are often on the edge of multiple fields of study and dealing with human-like intelligence. But it also means that there are a lot of other really interesting and equally advanced robotics projects out there, that fall under the shadow of their humanoid brethren. Simply because they are harder to explain and relate to (or fall short on the aesthetic factor).

Links & References

• Thanks to John for the inspiration
• Robonova soccer image courtesy of Robotgames
• Comments in the article How Do You Make a Robot Look Feminine?
• Rest came from my gray matter