4.12.2016

Multimillion dollar humanoid robot doesn't make for a good cleaner

Robotic humanoid butlers still have a way to go before you’ll be able to let them have free reign of the house, but Team IHMC from Florida and their multimillion dollar Google-developed US government Atlas robot are giving it a good try.The robotics team from the Florida institute for human and machine cognition taught the humanoid robot how to perform typical human
chores and housework, which proved more difficult than you might imagine, even for the team that came second in the 2015 Darpa Robotics Challenge.Atlas is a semi-autonomous system. The operator tells the robot where to be and what position to take, such as where to put its hands on a vacuum cleaner, and then the robot comes up with a plan of how to do that. For some chores Atlas’s actions are logical and human-like. Others require re-thinking of how to get the job done in a way that its quite different to the way a person would perform the action.John Carff, robot operator at IHMC told IEEE Spectrum: “It takes a lot of patience and out-of-the-box thinking to be a robot operator. When you approach a task or situation you’ve never seen before, you need to think of as many different ways of completing that task as you can and figure out what approach would be best for the robot. Many of the tasks Atlas performs are done a lot differently than a human would do the same task.”
A good example is the movement of using a jack to move a pallet. A human has the coordination to stop the jack rolling forward or backwards when pumping the handle to lift the pallet off the ground, but Atlas had to approach the task from the side and use its foot to stop the jack moving forward as it was raised.
For the team at IHMC, the whole exercise was a useful way of testing new algorithms and code beyond simply running the Robotics Challenge tasks over and over. For the rest of us, it proves that even the most advanced humanoid robots aren’t going to be a threat to our cleaning lives any time soon.

No comments:

Post a Comment