Automatic Detection of Driver Fatigue - Phase III

Image
Date Created
1999-06
Report Number
1999-30
Description
Sleep deprivation and sleep disorder continues to cause problems on the road. Reducing the number of accidents related to driver fatigue would save the society a significant amount of money and personal suffering. Monitoring the driver's symptoms can help determine driver fatigue early enough to prevent accidents due to lack of awareness. This report describes advances towards a non-intrusive approach for real-time detection of driver fatigue. It uses a video camera that points directly toward the driver's face and monitors the driver's eye to detect micro-sleeps, or short periods of sleep of about three-to-four seconds.

Simulation Validation: Evaluating Driver Performance in Simulation and the Real World

Image
Date Created
1998-07
Report Number
1998-28
Description
Simulation offers a cost-effective way to conduct research on collision avoidance and accident prevention. To be effective, simulated performance must be a valid measure of real world performance. This project sought to validate real world driving performance based on the performance of individuals driving in simulation. The study presents performance data on 14 male and 12 female volunteer subjects who drove a route adjacent to the University of Minnesota campus and then performed in a similar computer-generated driving route. Generally, subjects reported the simulated driving test comfortable and realistic; performance and characteristics of driving in the simulator closely paralleled the real world; the qualitative pattern of driving was similar; and errors and the control parameters of driving performance suggested acceptable reliability between both driving worlds. Researchers concluded that the simulator performed reliably and provided a valid set of performance data that could be used to better understand driving behavior, especially as it related to accident prevention and collision avoidance.

Human Factors Issues in Traffic Signing

Image
Date Created
1994
Report Number
94-33
Description
This work reports results of an experimental program on human factors issues in traffic signing. The first task examines the problems associated with the programming of signs for evaluation of driver response in simulation. It is concluded that growing technical tools permit traffic engineers to test proposed signage, and avenues of implementation are given. The second task examines driver response in simulation to multiple real-world signs. It is concluded that while much effort is given to distinguishing the utility of individual signs, multiple signs in combination produce more complex decrements. Recommendations are made as to maximum sign density. The final task provides an assessment of signage in future IVHS driving environments. It points to the role of signage as one component of communication. A list of issues for future signage implementation is given for consideration as the Department moves to provide safe and efficient transport for the people of Minnesota into the 21st century.

Following Advice from Traffic Advisories

Image
Date Created
1994
Report Number
94-29
Description
This report presents data using multiple-choice questionnaires to learn how drivers respond to traffic information in the form of advisory messages. Two experiments, comprising 112 participants, were conducted using the same technique and yielding similar results. The traffic information messages presented to participants varied in three respects; quantitativeness of information, imperativeness of advice, and timeliness of information. Two additional factors were examined; the amount of traffic congestion stated to be directly observable on the route and the stated accuracy of messages received in the past. Results obtained from the questionnaires indicate that the structure of the traffic message did influence the driver behavior. The propensity to depart from the planned route ahead of schedule was greater when respondents had; few exit options remaining, been told traffic levels were high, received accurate traffic information in the past, and had received messages which contained quantitative and/or imperative information. Traffic controllers with this knowledge of driver behavior could act to further reduce trip times and congestion by using the control tools currently available to them. The major conclusion we can draw from this study is that when possible and appropriate, advisory messages should contain accurate, timely, quantitative and imperative information.