Proactive driving safety evaluation: An evaluation of an automated traveler information system and investigation of hazard analysis data

The purpose of this dissertation was to assess the relative safety of a prototype in-vehicle automobile navigation device, and to evaluate consistency and meaning of data generated by the state-of-the-art in proactive safety evaluation techniques. Sixty subjects participated by driving an instrumented test vehicle equipped with Motorola's Mobile Navigation Assistant (MNA). Subjects were selected from younger (25-45) and older (over 65) age groups and varied in their experience using the MNA.; Data was collected while subjects navigated the test vehicle to fixed destinations using the MNA with and without auditory commands, and while using a conventional paper map and paper list of directions. Data concerning eye glance behaviors, driver performance measures, hazard analysis measures, subjective workload measures, and subjective opinions of safety were collected and analyzed from the experimental drives. The data generated from applying a hazard analysis technique was also evaluated and compared to the results of previous safety evaluation efforts.; The results of the statistical analyses show that there were some differences in eye glance behavior between navigation methods where the MNA conditions resulted in shorter glances away from the roadway but the glances occurred more often than with the paper map and direction list. However, increases in experience with the MNA tended reduce the amount of visual attention required. There were very few differences in driving performance and subjective workload measures indicating few differences in overall workload between conditions. Through subjective measurement, drivers consistently indicated the MNA was safer than conventional methods of navigation. The analysis of data generated by the hazard analysis technique showed it is consistent between studies and the relationship between driver errors, near misses, and accidents is roughly exponential in form.; No apparent differences in safety were found between conventional and computerized navigation methods when looking across all measures collected and analyzed. The analysis of data from the hazard analysis technique indicates it is a relatively robust method of safety evaluation and there might be a consistent relationship between accidents, near misses, and driver errors that could some day be used for accident prediction.