VTechWorks staff will be away for the Thanksgiving holiday starting Wednesday afternoon, Nov. 25, through Sunday Nov. 29, and will not be replying to requests during this time. Thank you for your patience.
Development of models for detection of automobile driver impairment
Dingus, Thomas A.
MetadataShow full item record
Two of the leading causes of automobile accidents are driver impairment due to alcohol and drowsiness. Apparently, a relatively large percentage of these accidents occur because drivers are unaware of the degree to which they are impaired due to these sources. The purpose of this research was to develop models which could detect driver impairment due to alcohol, drowsiness, or the combination of alcohol and drowsiness, and which could be practically implemented in an automobile. Such detection models, if successfully implemented in conjunction with a system to warn an impaired driver of his or her condition, could potentially save hundreds of lives each year. Six driver-subjects operated a computer controlled driving simulator during each of four conditions. The four conditions consisted of a control condition, an alcohol condition, a sleep-deprived condition, and a combination alcohol and sleep-deprived condition. Moderate levels of alcohol and sleep deprivation were used for this study. Nineteen performance and behavioral measures were collected during this study. Each measure was evaluated singly and in combination with other measures to determine potential value for detection of driver impairment. Detection models were then formulated using the most promising detection measures. The results indicated that a useful on-board drowsiness impairment detection device is possible and practical for highway driving. This device would also, in all likelihood, provide useful detection information regardless of whether low to moderate amounts of alcohol were present in a drowsy driver. The results also showed that on-board alcohol impairment detection may be possible at moderate to high BAC.