Fused quad audio/visual and tracking data collection to enhance mobile robot and operator performance analyses

摘要

Collecting accurate, adequate ground truth and experimental data to support technology evaluations is critical in formulating exact and methodical analyses of the system's performance. Personnel at the National Institute of Standards and Technology (NIST), tasked with developing performance measures and standards for both Urban Search and Rescue (US&R) and bomb disposal robots, have been designing advanced ground truth data collection methods to support these efforts. These new techniques fuse multiple real-time streams of video and robot tracking data to facilitate more complete human robot interaction (HRI) analyses following a robot's experiences. As a robot maneuvers through a test method, video and audio streams are simultaneously collected and fed into a quad compressor providing real-time display. This fused quad audio/visual data provides a complete picture of what the operators and robots are doing throughout their evaluation to not only enhance HRI analyses, but also provide valuable data that can be used to aid operator training, encourage implementation improvements by highlighting successes and failures to the developers/vendors, and demonstrate capabilities to end-users and buyers. Quad data collection system deployments to support US&R test methods/scenarios at the 2007 Robot Response Evaluation in Disaster City, Texas will be highlighted.