Dynamic motion capture footage of humans is tracked using an accurate new physics-based prior.

Project to create an online video game interface with which users can remotely control robotic dogs to play soccer. The purpose of the project is to gather learning data for the dogs, so they can eventually be taught to play the game on their own.

A color-tracking romance triangle between 3 Roombas.

Several interaction modalities are combined to allow natural teaming with an iRobot packbot. Person following, gesture recognition, and speech recognition are used to create a platform which is robust to indoor and outdoor environments.

Robot learns soccer task (goal-seeking) from teleoperated demonstration.

Sparse control of the DLR robot hand using the 2D screen for user input.

Teleoperated control of the NASA Segwanaut, a Robonaut on a Segway base, using an inertial motion capture system, described in our Humanoids 2004 paper.

Final match of CS148 Roomba/Create 2-on-2 robot soccer competition.