Submitted by Merreton: Remember the shape-shifting cyborg that chased Arnold Schwarzenegger through Terminator 2? You know the one: the robot from the future that can change forms at a moment's notice. One minute, he's Robert Patrick, the next he's mercury. (Okay, polymimetic alloy.) To most people this villainous character—known as the T-1000—is an awfully cool special effect. When James Cameron's sci-fi blockbuster hit movie screens in 1991, the T-1000 was the ultimate in computer-generated imagery, and it's still hailed as a major turning point in the history of motion pictures. To Wei-Min Shen, it's just a blueprint. He wants to build the real thing. As the director of the Polymorphic Robotics Laboratory at the University of Southern California, Dr. Shen spent the past year building what he calls superbots, modular automatons that can break themselves apart and re-form as entirely new devices capable of entirely new things. Yes, he's a long way from building a cyborg that can turn itself into molten metal and slip through the bars of a jail cell, Robert Patrick–style. But he's just another 12 months from finishing a version that can run for a full kilometer, then reshape itself for a climb up a 60-degree slope. "Our dream is to build a robot that changes shape and size by itself, depending on the environment and the task at hand," Shen says. "We hope to build a robot that can become, say, a snake if it needs to crawl through a narrow space. Then, if it needs to climb some other obstacle, it can grow legs. Or, if it needs to go downhill, it can change into a ball and roll. We often say that we're building a baby version of the T-1000." You can see prototypes in action on the lab's Web site. At this point, the demos are pretty rudimentary—the bots can't actually reconfigure themselves yet; they need a little help from researchers. But the basic building blocks are there. Today, a superbot consists of up to six separate modules capable of working in tandem. When connected in a ring, they can roll across the floor. In another form, they can crawl. They can even shovel sand from a sandbox or pour Dr. Pepper into an empty cup. Measuring about 2 inches deep and 4 inches long, each individual module is built around an ultra-minicomputer with a Atmega chip and 256KB of memory. Using a trio of tiny motors and various sensors that determine where it is and which way it's facing, this tiny PC can move the module in three distinct ways. But the movement of each module is dependent on what all other modules are doing. All the while, they're communicating via miniature radio transmitters and receivers.
View: Full Article | Source: ABC News