Servo - February 2010

LEARNS FROM ACTIONS

The ability to teach robots new tasks is rather important. But even more important is teaching robots to generalize one thing or task and apply it to many different variations on that thing or task. Recall the ASIMO that learned to generalize different types of objects, for example, show it a chair and ASIMO could figure out what other objects were chair-like.

PR2 from Willow Garage has taken this concept of generalization a step further by applying it to actions, not just things. When you teach PR2 an action (by providing an example of the action), it can understand the fundamental components of the action and apply them to different variations of the action. So, if you teach PR2 to pour a beverage out of a can into a cup, it doesn’t matter what kind of can or what kind of cup, and more importantly, where they are relative to each other or if you move the can and/or cup around in the middle. This is huge for robots to be able to do, because having to teach a robot an action for “pour from can into cup on counter” and then a separate action for “pour from can into cup on low table” isn’t just tedious, it’s one of those things that is going to bring development to a screeching halt.

There are a lot of major obstacles to surmount when it comes to creating a robot that can be practical and useful around people who know nothing about robots … from stuff like plugging into a normal outlet to recharge, to not squeezing things too hard.

HOT SHOTS Scientists from the Woods Hole Oceanographic Institution managed to capture a volcano in action in the Pacific, about 4,000 ft. below the surface with the help from Jason — a remote controlled underwater robot. Photos and videos show that what is spewing out of the West Mata volcano is boninite lava which is thought to have only been present on extinct volcanoes. Interestingly enough, the team managed to find some shrimp in the same area and will be studying their DNA.