Quantcast

Human Touch Makes Robots Smarter

November 5, 2013

In this video, the robot does not know how to move a knife safely and learns through Interactive zero-G and tablet touch user interaction on Baxter for grocery checkout application. With high degree-of-freedom manipulators, one needs to find a trajectory that is not only valid from a geometric point-of-view (i.e., feasible and obstacle-free, the criterion that most existing planners focus on), but also satisfy the user’s preferences. In this work, we propose an algorithm to learn such preferences via eliciting online feedback from the user, which does not need to be an optimal demonstrations.

credit: Cornell University Personal Robotics Lab / PhD student Ashesh Jain / Prof. Ashutosh Saxena.



comments powered by Disqus