(www.plasticpals.com) Rodney's Robot Revolution is an Australian documentary from 2008 about former MIT Professor Rodney Brooks' attempt to build a Go-playing robot for the U.S. military. Go is a traditional board game similar to chess or checkers, where two players take turns placing black or white stones on a board. Besides recognizing the game board and rules, it would have to possess delicate manipulation in order to pick and place the small pieces without disturbing the others. And Brooks and his team at MIT would have to build this robot from scratch on a small budget in just five months.
It should be possible, given that a small team of researchers in China have built one for the same purpose, on a similarly tight budget, at around the same time. However, their Go-playing robot is cutting corners. As you can see in the photo below, the board has pits for the pieces to fall into, and it appears to have a dispenser for the pieces rather than a bowl, like the human. Brooks' planned robot won't have either of these cheats to help it out. A Chinese Go-playing Robot
We're given a primer on Rodney's work over the past 25 years, from his impatience with the Stanford Cart in the '70s, to the founding of his company iRobot in the '90s. He's credited with revolutionizing artificial intelligence through a "bottom-up" approach, exemplified by his hexapod Genghis, which combined sensing and action without the planning. Along the way we're introduced to Domo, one of the cognitive robots developed at MIT. It has a head with two bright blue eyes with which it can recognize faces, and a pair of arms with simple hands to grasp objects. It's the result of seven years of research, and is distinguished by its compliant arms which make it safe to interact with.
In comparison, the Go-playing robot isn't too exciting; it's a rather bulky and inelegant contraption. But Brooks sees a future for this sort of robot in the manufacturing sector, where precise manipulation jobs are often outsourced where ever the labor is cheap. His company, Heartland Robotics, is busy working on just such a robot. He's not alone, and will face stiff competition from the likes of ABB's Frida, Yaskawa's Motoman, and Kawada Industries' Nextage.
Brooks then travels to Japan to meet with Osaka University's Prof. Asada, one of the luminaries working in cognitive robotics through the bottom-up approach. In this segment, Japan's acceptance of robots as potential partners is explained by cultural differences (animism) and pop cultural icons (Astro Boy). In Japan, robots are being developed to one day assist with the country's rising elderly population, but Georgia Tech's Ron Arkin is skeptical. He suggests that robots can't replace human care takers since they lack the human touch. We've all heard of the horrible abuses that have taken place at retirement homes run entirely by humans, so couldn't robots help to alleviate stress (rather than outright replacing them)?
Though Brooks seems quite taken by his encounter with CB², he rightly criticizes the ASIMO puppet shows put on at science museums. These demonstrations, controlled entirely by human operators hidden from view, give people the impression that robots are smarter than they really are. On the other hand, he doesn't seem to have a problem with Kismet, which is guilty of feigned intelligence too (as pointed out by an anthropologist). To label humanoids like ASIMO mere "toys" smacks of arrogance, or perhaps sour grapes. So what if Honda's focus has been on the mechanical and not the intellectual side of the argument? Both are important, even if artificial intelligence is the more difficult problem to solve.
It should be possible, given that a small team of researchers in China have built one for the same purpose, on a similarly tight budget, at around the same time. However, their Go-playing robot is cutting corners. As you can see in the photo below, the board has pits for the pieces to fall into, and it appears to have a dispenser for the pieces rather than a bowl, like the human. Brooks' planned robot won't have either of these cheats to help it out. A Chinese Go-playing Robot
We're given a primer on Rodney's work over the past 25 years, from his impatience with the Stanford Cart in the '70s, to the founding of his company iRobot in the '90s. He's credited with revolutionizing artificial intelligence through a "bottom-up" approach, exemplified by his hexapod Genghis, which combined sensing and action without the planning. Along the way we're introduced to Domo, one of the cognitive robots developed at MIT. It has a head with two bright blue eyes with which it can recognize faces, and a pair of arms with simple hands to grasp objects. It's the result of seven years of research, and is distinguished by its compliant arms which make it safe to interact with.
In comparison, the Go-playing robot isn't too exciting; it's a rather bulky and inelegant contraption. But Brooks sees a future for this sort of robot in the manufacturing sector, where precise manipulation jobs are often outsourced where ever the labor is cheap. His company, Heartland Robotics, is busy working on just such a robot. He's not alone, and will face stiff competition from the likes of ABB's Frida, Yaskawa's Motoman, and Kawada Industries' Nextage.
Brooks then travels to Japan to meet with Osaka University's Prof. Asada, one of the luminaries working in cognitive robotics through the bottom-up approach. In this segment, Japan's acceptance of robots as potential partners is explained by cultural differences (animism) and pop cultural icons (Astro Boy). In Japan, robots are being developed to one day assist with the country's rising elderly population, but Georgia Tech's Ron Arkin is skeptical. He suggests that robots can't replace human care takers since they lack the human touch. We've all heard of the horrible abuses that have taken place at retirement homes run entirely by humans, so couldn't robots help to alleviate stress (rather than outright replacing them)?
Though Brooks seems quite taken by his encounter with CB², he rightly criticizes the ASIMO puppet shows put on at science museums. These demonstrations, controlled entirely by human operators hidden from view, give people the impression that robots are smarter than they really are. On the other hand, he doesn't seem to have a problem with Kismet, which is guilty of feigned intelligence too (as pointed out by an anthropologist). To label humanoids like ASIMO mere "toys" smacks of arrogance, or perhaps sour grapes. So what if Honda's focus has been on the mechanical and not the intellectual side of the argument? Both are important, even if artificial intelligence is the more difficult problem to solve.
