In his position, Brooks says that he is “way too close” of a dexterity robotics figure humanoid when it fell a few years ago. He has not dared to approach one since. Even in promotional videos of humanoid companies, Brooks notes, people are never shown close to moving humanoid robots, unless separated by furniture, and even then the robots only shake minimal.
This safety problem extends beyond Accidental Falls. For humanoids to play their promised role in health care and factory institutions, they need certification to work in zones that are shared with people. Current running mechanisms make such a certification virtually impossible under existing safety standards in most parts of the world.
The Humanoid Apollo robot.
Credit: Google
Brooks predicts that within 15 years there will indeed be many robots called “Humanoids” that perform different tasks. But ironically they will not look nothing on the two -way machines of today. They will have wheels instead of feet, varying numbers of arms and specialized sensors that do not show like human eyes. Some will have cameras in their hands or look down their abdominal muscles. The definition of “humanoid” will shift, just like “flying cars” now means electric helicopters instead of can -can -it -through aircraft, and “self -driving cars” means vehicles with remote human monitors instead of really autonomous systems.
The billions that are currently being invested in the forcing of the rigid, only-vision-all-humanoids of today to learn agility will largely disappear, Brooks argues. Academic researchers are making more progress with systems that contain touch feedback, such as MIT's approach using a glove that transmits sensations between human operators and roboths. But even these progress remain far from the extensive touch detection that makes human agility possible.
Nowadays, few people spend their days near humanoid robots, but the three-meter rule of Brooks stands as a practical warning for challenges from someone who has built these machines for decades. The gap between promotional videos and deployable reality remains large, not only measured in years, but also in fundamental unsolved problems of physics, detection and safety.
