Check out the Kuratas mecha droid, which is technically an art project that just had its first public showing at Wonder Fest 2012. The 4-meter tall, 4.5-ton droid may be designed purely for show, but it is also smart: It has a smartphone-based touchscreen remote control interface and also uses a Microsoft Kinect to quickly image the mecha pilot's gestures to control the mecha's arms. It's silly and meant to be fun, but the faux-gattling guns on the droid do remind us that a battle-ready droid like this isn't too far from being real, particularly when you recall Petman.
Hexapedal robots are pretty common in the robot world—the feet give a robotic platform extra stability on uneven terrain where robots with two feet, four feet or even tracks may have difficulty navigating (even NASA's robot Mars explorer, due to land on the red planet imminently, has six leg-like appendages that end in a wheel). But a hexapod robot from UC Berkeley and the University of Pennsylvania, shown on the Automaton blog, adds a new and weirdly effective twist to the design: X-RHex Lite has a long tail as well as six wheel-like legs. The idea of the powered tail is that, much like a cat, if it falls far it can orientate itself in mid-air so it lands upright, and can help it keep its balance when moving fast across unusual obstacles. Tricks like this will be useful for future robot explorers, military machines or search-and-rescue robots.
Hanson Robotics machines are always interesting thanks to their advanced programming and highly expressive faces, partly due to a novel material called flubber. The company has just revealed its latest edition of its education/experimental robot Zeno. He's a 27-inch tall android with up to 37 degrees of freedom, up from the previous 33, and the video hints that when he arrives in 2013 he may even be cheaper than the many thousands of dollars the current Zeno costs.
Robots with sensitive skin. Giving robots a sense of touch is more important than you think for helping them pick up difficult objects and even avoid injuring soft and fragile humans, so new work by South Korean researchers is very welcome. Scientists at Seoul National University have worked out how to weave polymer nanohairs into a touch sensor that could let robots have a sensitive enough surface to detect a human heartbeat—the 50 nanometer thick hairs rob when they're moved and generate tiny electrical forces that the system identifies as pressure, or different types of sliding and twisting contact.
MIT's shiny spheres. We've known about MIT's simple but powerful Spheres project for some time—it's flying aboard the ISS as a test-bed for future robot swarms that may be used to perform experiments or even to inspect or fix the exterior of a spacecraft. Now MIT's revealed that the Spheres are going to get two big upgrades: The first is a powerful vision system called Vertigo that uses twin cameras to deliver stereoscopic vision—potentially useful for telepresence remote-control, and also able to assist with autonomous vision-based movements. The second, Rings, are large electromagnetic coils that encircle each bot—they can be used to group individual robots together, form precise formations and they could also act as wireless power transfer.
Gripping drones. If the idea of numerous flying drones, owned by the military or police or other government bodies already makes you nervous you won't like this: Researchers at Drexel University have suggested that future drones may benefit from dangling robotic arms, and have developed some prototype designs. The idea is that a low-hovering drone could pick up objects in a search-and-rescue situation or even help in construction tasks.
A few weeks back we posed a question in This Week In Bots that may turn out to be important in our near-future economy: What if robots took on more people's jobs? In a sense it's partly inevitable because robots are technically designed to do precisely this thing, and they're actually more precise and reliable themselves—meaning the costs of production.
Now a post on ZDNET leads to a more sophisticated question that really complicates the matter: What if we created workplaces that were entirely robotized, not even designed for humans to enter, and which may even be sealed off so no one can go inside?
It's not a ridiculous issue at all, and ZDNET points out that one of the first situations this might happen is inside datacenters. These are already pretty human-hostile environments, with enormous stacks of machinery, noise, a mass of wiring, sophisticated air conditioning and so on. They are also designed so that people can get among the machinery to access broken equipment, and organizing a datacenter like this may actually not be the most efficient way to do so in terms of electronic efficiency (or for maximized cooling).
Thus it's entirely plausible that near-future datacenters could be peopled with robot arms to extract failed drives or electronics racks, and that there wouldn't even be space for a person to get inside—unless in an extreme emergency. A similar model to this already exists in Amazon's warehouses, if you think about it, with the automated Kiva robots picking up and moving stock around all by themselves. If these warehouses were fully robotized, then it could simplify their design as well as bump up their efficiency.
Facebook even suggested to ZDNET that a robotized version of their clever datacenters could actually be welded shut and dealt with as a "degrade and replace" unit—where you simply waited for the system to fail to the point of being inefficient, before replacing it with a new one and opening the old one to harvest useful hardware.
It all makes excellent sense. But there's actually a difficult philosophical issue at the heart of the idea. If we build human-hostile workplaces, we're ceding an extraordinary amount of control to robots, and there may be inherent physical risks if humans did need to go inside (cue any number of sci-fi movies where action happens in a highly robot-populated factory). Then at some point, a robotic factory will be churning out robots to work in other factories, and we may even find it useful to let the robot system work out its own optimal working patterns, relying on its ability to learn what works best.
To some human minds, and in the political sphere this would be a terrible—perhaps even terrifying—state of affairs, and result in humans deprecating themselves from many jobs because our robotic tools are just much more efficient at the tasks we need them to do.
[Image: Flickr user jurvetson]