Why Humanoids?
By Robotics Trends Staff - Filed Mar 10, 2006
I would like you to try the following mind experiment. It is nothing on par with the mind experiments involving Einstein’s elevator or Schrödinger’s cat, but it does provide some insight into a debate within the robotics community, and a significant debate at that.

To set the stage, picture yourself viewing representations of robots, maybe as photos in an article, on television, as a film clip on the Web, or better yet, a live demonstration. Now ask yourself the following question, and try not to think too hard before you answer. When viewing robots, what type or form factor draws your attention the quickest and holds it the longest? More specifically, what if the choice was between an industrial robot and its humanoid counterpart? OK. Now let’s try a robotic vacuum cleaner and a humanoid. How about a surgical robot and a humanoid, or a lawn mower and a humanoid? You get the point.

imageI think you will find that there is little disagreement that humanoid robots hold a distinct fascination across all age groups and levels of technical literacy. To put it more plainly… the humanoid robots win hands down in popularity contests. I have seen this scenario played out many times at the various robotics events I have attended or produced. Consider the RoboCup championships held in Osaka, Japan in July 2005. The RoboCupRescue and RoboCupJunior events, along with the various RoboCupSoccer competitions, were all well attended by large numbers of enthusiastic fans. However, it was the humanoid RoboCupSoccer events that were the biggest crowd pleasers, even though these events had the least amount of real ‘action’.

Other examples abound. Without going into the details just consider R2D2, Honda’s Asimo, NASA’s Robonaught, or the success of Wow Wee Robotics’ Robosapien. What about the DARPA Grand Challenge you might ask? Surely, an event where 23 autonomous robotic vehicles race over a 131-mile desert course for a $2M prize captures the imagination? Undoubtedly. The event received wide press coverage. However, imagine if the Grand Challenge was replaced with a RoboDash event – 10 humanoid robots lined up and competing in a 100 yard dash. Now try to visualize the amount of press that event would receive. It would be staggering. Does the research and results of the Grand Challenge have greater applicability in the real world than our fictional humanoid RoboDash? Sure. Does it matter to the world at large? Not one bit.

In Broad social terms, the technical objections to the development of humanoid robots are moot.
The emotional appeal of the humanoid form factor simply trumps other concerns. Research and
resources will continue to be dedicated to humanoid robotics at the university level, while toy and
educational robotics producers will grow the market from the “bottom up” by releasing increasingly
more powerful and complex robots.

imageWhile there is no disagreement as to the emotional appeal of humanoid robots, the same can not be said as to the role of humanoid robots or the humanoid form factor for serious work. The arguments against devoting large amounts of resources, in terms of time, expertise and money, to the development of humanoid robots, are numerous and varied. Most, however, focus on the fact that the humanoid form factor is not an optimal platform in terms of functionality, performance and cost.

The technical hurdles to overcome to develop humanoid robots are substantial. Locomotion and power utility come to mind, but they are really only the beginning. We are currently at the stage of robotics development where most products, whether they serve the military, civil, commercial or consumer markets, are devoted to a single task, or at the very most, serve a limited purpose. The next step is to provide these simple robots with the ability to interoperate with other simple robots in networked teams, to perform more complex tasks such as serving as mobile sensor networks.

Yet, such networked systems will not be released from the lab for years to come. The robotics industry is simply not at the point where autonomous, multitasking functionality can be realistically deployed. You will notice that I have said nothing about natural language interfaces or emotive interactions. Yet this is exactly the type of functionality that the ‘humanoid’ in humanoid robotics implies.

In broad social terms, the technical objections to the development of humanoid robots are moot. The emotional appeal of the humanoid form factor simply trumps other concerns. Research and resources will continue to be dedicated to humanoid robotics at the university level, while toy and educational robotics producers will grow the market from the ‘bottom up’ by releasing increasingly more powerful and complex robots. In this way there is a type of pincer movement occurring, where the end result will be functional, autonomous, humanoid robots. Notice I did not say when these humanoid wonders will enter our lives. Think of humanoid robotics as you would a manned mission to Mars. It will take a great deal of work, money and time, and in no way will we be compensated equally in terms of spinoff technology. It could also be argued that the money could have been better spent. But will there be a manned mission to Mars? Absolutely.

humanoid robotI understand and agree with most of the arguments against the amount of emphasis placed on humanoid robotics development, especially if it comes at the expense of efforts that will result in real products, solving real problems, in the short term. But there also seems to be an undercurrent of disapproval, often expressed as heated verbal opinions, directed at those who take the humanoid route, especially if those robotic developers come from multinational corporations and make large public displays of their humanoids.

Much of the criticism centers on the large amount of press dedicated to humanoid robots that, speaking frankly, do very little. The high end Japanese robots from the likes of Honda and Toyota run through carefully scripted demonstrations supported by scores of engineers. On more than one occasion I have personally heard such performances compared to Professor Marvel in the Wizard of OZ manipulating controls behind a curtain to create a breathtaking display of wizardry for an astounded audience. What appears to be miraculous autonomous behavior is nothing more than preprogrammed sequential actions, supported with a dash of remote control. Brilliantly engineered and beautifully choreographed remote control, but remote control nonetheless.

Such large public displays of humanoid robotics - and just to be clear we are talking about Japanese humanoids here - are dismissed by many as simply a marketing ploy - a high-end, high-touch, way of exhibiting engineering excellence. An equal number of people write off the displays as demonstrations of national pride, or at least national ego (Japan is universally acknowledged as the home of the world’s most stunning humanoid robots).

People will continue to develop humanoid robots because eventually there will be
humanoid robots. It is a classic case of circular reasoning, but true nonetheless.

Many others believe that the emphasis of humanoid robots in Japan is the direct result of a Japanese national preoccupation with comic books, and one comic book character in particular. That character, perhaps cultural icon is a better term, is Astro Boy, or Mighty Atom as he is known in Japan, a humanoid robot from a cartoon series which first screened in black and white on Japan’s Fuji Television network in 1963. According to Toru Takenaka, Honda’s chief engineer for the ASIMO project, “I belong to the Atom generation. When I was a child, I loved Atom and Tetsujin 28 (another cartoon robot), and I used to be immersed in the robot world.”

The impact of Astro Boy on Japanese roboticists cannot be overstated, but nor is it as strange as it seems at first blush. Popular culture has had a significant impact on the robotics world, and that impact is not limited to robotic toys. For example, Helen Granier, iRobot’s Co-founder and Chairman of the Board, has often stated that it was R2D2 that first engendered in her an interest in robotics.

imageSo let me get this straight. Japan’s preoccupation with humanoid robots, and the spending of considerable sums developing some of the more famous humanoid robots on the planet, is part of a worldwide marketing and branding campaign (a good business investment) and part fixation with a 40 year old cartoon character (a bad business decision). These robots are funded and developed as national representatives of engineering excellence, even though they have little chance of delivering benefits at the national level and in the near future. Unlike, say, the bullet train, another photogenic engineering feat and source of national pride that actually provides a useful service.

Given the incongruities and inconsistencies listed above, I would say that the emphasis on humanoid robots in Japan and in other quarters is not the result of national pride or homage to a cartoon character taken to an extreme. Nor is it naivety, lack of business acumen, or a 20 year old branding campaign. But what then?

In such situations I find it useful to invoke Occam’s razor, the principle attributed to the 14th-century English logician and Franciscan friar, William of Occam, who wisely noted that when one is confronted with multiple predictive theories, one should make no more assumptions than needed. That is, all things being equal, choose simpler explanations over more complicated arguments. In the case of humanoid robotics, there is a straightforward and simple answer to the question “Why humanoids”?

The easy answer is that robots will eventually operate in our home and workplaces, so their form factor must conform to those of humans. But I dismiss this explanation as well. It is possible, and probably easier, to build robots that can navigate stairs, doorways and halls, but that are not necessarily humanlike in appearance. No, people will continue to develop humanoid robots because eventually there will be humanoid robots. It is a classis case of circular reasoning, but true nonetheless.

Dan Kara is President of Robotics Trends, the producer of the RoboBusiness (http://www.robobusiness2006.com) and RoboNexus (http://www.robonexus.com) conferences, and publisher of Robotics Trends (http://www.roboticstrends.com), and online news, information and analysis portal covering the personal, service and mobile robotics market. He can be reached at dk(at)roboticstrends.com.

This article was originally published in the January 2006 issue of Servo Magazine.

Much of the criticism centers on the large amount of press dedicated to humanoid robots that, speaking frankly, do very little. The high end Japanese robots from the likes of Honda and Toyota run through carefully scripted demonstrations supported by scores of engineers. On more than one occasion I have personally heard such performances compared to Professor Marvel in the Wizard of OZ manipulating controls behind a curtain to create a breathtaking display of wizardry for an astounded audience. What appears to be miraculous autonomous behavior is nothing more than preprogrammed sequential actions, supported with a dash of remote control. Brilliantly engineered and beautifully choreographed remote control, but remote control nonetheless.

Such large public displays of humanoid robotics - and just to be clear we are talking about Japanese humanoids here - are dismissed by many as simply a marketing ploy - a high-end, high-touch, way of exhibiting engineering excellence. An equal number of people write off the displays as demonstrations of national pride, or at least national ego (Japan is universally acknowledged as the home of the world’s most stunning humanoid robots).

People will continue to develop humanoid robots because eventually there will be
humanoid robots. It is a classic case of circular reasoning, but true nonetheless.

Many others believe that the emphasis of humanoid robots in Japan is the direct result of a Japanese national preoccupation with comic books, and one comic book character in particular. That character, perhaps cultural icon is a better term, is Astro Boy, or Mighty Atom as he is known in Japan, a humanoid robot from a cartoon series which first screened in black and white on Japan’s Fuji Television network in 1963. According to Toru Takenaka, Honda’s chief engineer for the ASIMO project, “I belong to the Atom generation. When I was a child, I loved Atom and Tetsujin 28 (another cartoon robot), and I used to be immersed in the robot world.”

The impact of Astro Boy on Japanese roboticists cannot be overstated, but nor is it as strange as it seems at first blush. Popular culture has had a significant impact on the robotics world, and that impact is not limited to robotic toys. For example, Helen Granier, iRobot’s Co-founder and Chairman of the Board, has often stated that it was R2D2 that first engendered in her an interest in robotics.

imageSo let me get this straight. Japan’s preoccupation with humanoid robots, and the spending of considerable sums developing some of the more famous humanoid robots on the planet, is part of a worldwide marketing and branding campaign (a good business investment) and part fixation with a 40 year old cartoon character (a bad business decision). These robots are funded and developed as national representatives of engineering excellence, even though they have little chance of delivering benefits at the national level and in the near future. Unlike, say, the bullet train, another photogenic engineering feat and source of national pride that actually provides a useful service.

Given the incongruities and inconsistencies listed above, I would say that the emphasis on humanoid robots in Japan and in other quarters is not the result of national pride or homage to a cartoon character taken to an extreme. Nor is it naivety, lack of business acumen, or a 20 year old branding campaign. But what then?

In such situations I find it useful to invoke Occam’s razor, the principle attributed to the 14th-century English logician and Franciscan friar, William of Occam, who wisely noted that when one is confronted with multiple predictive theories, one should make no more assumptions than needed. That is, all things being equal, choose simpler explanations over more complicated arguments. In the case of humanoid robotics, there is a straightforward and simple answer to the question “Why humanoids”?

The easy answer is that robots will eventually operate in our home and workplaces, so their form factor must conform to those of humans. But I dismiss this explanation as well. It is possible, and probably easier, to build robots that can navigate stairs, doorways and halls, but that are not necessarily humanlike in appearance. No, people will continue to develop humanoid robots because eventually there will be humanoid robots. It is a classis case of circular reasoning, but true nonetheless.

Dan Kara is President of Robotics Trends, the producer of the RoboBusiness (http://www.robobusiness2006.com) and RoboNexus (http://www.robonexus.com) conferences, and publisher of Robotics Trends (http://www.roboticstrends.com), and online news, information and analysis portal covering the personal, service and mobile robotics market. He can be reached at dk(at)roboticstrends.com.

This article was originally published in the January 2006 issue of Servo Magazine.

<< Return to story