Situation awareness (SA) is critical to the successful operation of an unmanned vehicle. Since 2003, we have studied numerous robot control systems intended for emergency response. We have designed evaluation methods and produced design guidelines for HRI. From this, we implemented our own multi-phase video-centric USAR system. Our platform is an iRobot ATRV-Jr featuring a 2.8GHz Pentium IV processor, a full sonar ring (26 sonars), two Canon VC-C4 pan-tilt-zoom cameras (front and rear), a forward- facing SICK laser, a forward-facing FLIR Systems Indigo A10 camera, a carbon dioxide sensor, an audio speaker and microphone, and a lighting system. The robot currently uses four levels of autonomy, ranging from tele-operation to a goal-centric mode.

Our interface utilizes many novel features such as the fusion of directsensor information to increase SA. Specifically, we combine laser and sonar information to indicate obstacles within a one-meter radius of the robot (shown directly below main video display) and map the environment during operation (shown upper left of main video display). We also fuse FLIR with video (shown in the main video display), which allows the operator to easily notice potential unseen victims.

Other innovative features are included in our implementation. On the main video display, a cross-hair indicates the current orientation of forward camera. A small mirrored video display simulates a rear-view mirror in the upper right. We have the ability to swap front and rear camera views; the drive controls and distance information are remapped to reflect this change. This virtual exchange eliminates the need for the robot to backup, which results in more efficient operation.