Sagittarius: A Human-Assisted UAV/UGV Team for Cooperative Tracking of Elusive Dismounts

iRobot Corporation and Carnegie Mellon University (CMU) propose to develop the Sagittarius UAV/UGV team for tracking elusive human dismounts in complex urban terrain. iRobot will develop the UGV; CMU will develop the UAV; and iRobot and CMU will jointly develop techniques to enable the UGV and UAV to cooperate. The UAV will be equipped with multispectral cameras on a gimbaled, downward-looking mount. Active vision algorithms developed by CMU will separate the target from the background. Predictive tracking algorithms will determine the optimal path for the UAV to minimize occlusions between the UAV and the target. The UGV will be equipped with stereo FLIR cameras on a pan/tilt base that will enable the UGV to track human targets based on their temperature, size, and a target tracking vector supplied by the UAV. iRobot will use its extensive experience with behavior-based robotics to develop a control system to enable the UGV to robustly track elusive humans. During Phase I, we will demonstrate the UGV thermal vision system tracking a human target, the UAV active vision system tracking a human target from the air, and simulated UAV/UGV coordination techniques. We will also design the Phase II prototype UAV/UGV system.