Leverage NIST knowledge and capabilities in task analysis, perception, autonomous vehicle systems, and performance measurements to support an FWHA project to evaluate visibility requirements for driving two-lane country roads at night.
The project calls for the Texas Transportation Institute and SAIC to analyze human drivers, and to evaluate driver performance on a closed course. NIST applies 4D/RCS principals to analyze the driving task, and evaluates and applies vision-based perception algorithms for implementation on the NIST HMMWV testbed, to enable evaluation of machine performance on the same tasks on the same closed course. FHWA expects to be able to learn more about the visibility requirements by examining both human drivers, and autonomous vehicle algorithms, especially with respect to perception and decision making. The scenario calls for the use of color camera information, acquired through foveal and peripheral cameras approximating human vision resolution. The goal is improved understanding of the role and characteristics of roadway markings (center and edge lines) and other indicators (post-mounted delineators, etc.) so that improvements in safety can be achieved. The standards for such features are overdue for updating, having been developed prior to the time of modern vehicles.
Lead Organizational Unit:el
Mike Shneier, Project Manager
100 Bureau Drive, M/S 8230