Fun Stuff

A playlist of videos showing an emergent group mind on swarms of 160-260 Kilobot robots learning to recognize and react to images can be found here. The following are three videos from that list:

In the first video (above) the swarm detects a peace sign and so creates a smiley face.

In the second video (above) the swarm detects a biohazard sign and so creates a frowny face.

More information on this research can be found here.

A playlist of videos showing the RRT-X motion planning/replanning algorithm in various spaces can be found here. The following are two videos from that list:

In the first (above), a holonomic robot is moving among unpredictable obstacles. The robot assumes that the robots will continue straight along their current trajectories---whenever this assumption turns out to be false, then the robot replans. The video show the real-time reaction of the robot. Colored lines represent the estimated paths of the robot/obstacles into the future. For visualization purposes nodes have been projected into the spatial dimensions of the state space.

In the second, a Dubins robot is provided with an inaccurate map of the world and must replan whenever its sensors detect that a new obstacle has appeared or an old obstacle has disappeared. The video show the real-time reaction of the robot. Color represents the distance to the goal, the search tree edges are drawn as straight lines between the nodes (not as the Dubins trajectories the robot will actually follow). For visualization purposes nodes have been projected into the spatial dimensions of the state space.

---

The inexpensive Prairiedog platform that I helped to develop (ROS Wiki Documentation):

Here is a video that Aaron Cepher and Ilya Kushnir made showing wireless control of Prairiedog using neural responses read by an electro-encephalogram headset (in other words, using brain waves to control the robot):

In Fall 2009 I was T.A. for the Multi Robot Systems Class, check out what a few of the students put together using their own system (and our localization sensors):