[Tyler]

Gesture-based Human Robot Interaction
     


According to a Brown
University news release, a DARPA-funded team of researchers at the
Brown Robotics Lab have
modified an iRobot Packbot to interact with humans based on both verbal
commands and gestures. The Packbot is equipped with a CSEM SwissRanger SR300
depth camera for vision. The researchers developed software that
could recognize humans in the scene as distinct objects. Once the robot
identifies its human "commander" the robot tries to maintain a specific
distance to that person and observe their gestures. This allows the
robot to act as a team member, silently follow gestural cues of the
leader, just like human members of an military team. The video above
shows the robot interacting with humans who communicate with it by
making gestures. For more, see the
Brown website on Mobile
Human-Robot Teaming Enabled by Depth Imaging. You can find all the
technical info in the paper recently published by the team, Mobile
Human-Robot Teaming with Environmental Tolerance (PDF format)
   

http://robots.net/article/2788.html