Driver education classes were once seen as a remedy for young drivers' overinvolvement in crashes, but research results from the early 1970s were disappointing. Few changes in the content or methods of instruction occurred until recently, but this could change rapidly. Personal computers (PCs) can now present videos or photorealistic simulations of risky, cognitively demanding traffic scenarios that require quick responses without putting the participant at risk. As such programs proliferate, evaluating their effectiveness poses a major challenge. We report the use of a fixed-base driving simulator to study the effects of both experience on the road and PC-based risk awareness training on younger drivers' part-task simulator driving performance in risky traffic scenarios. We ran three groups of drivers on the simulator: one group first trained on the PC (younger, inexperienced drivers) and two groups who received no PC training (younger, inexperienced and experienced drivers). Overall, the younger, inexperienced drivers who were trained on a PC operated their vehicles in risky scenarios in ways that differed measurably from those of the untrained younger, inexperienced drivers and, more important, in ways that we believe would decrease their exposure to risk considering that, on average, their behavior was more similar to the behavior of the untrained, experienced drivers. More research is needed to demonstrate whether these findings apply on the open road to the larger population of younger drivers. However, at least initially, the research suggests that PC-based risk awareness training programs have the potential to reduce the high crash rate among younger, inexperienced drivers.

No abstract

Summary: The Transit Cooperative Research Program of the TransportationResearch Board recently sponsored an 18-month research program to develop a set of Guidelines that transit agency trainers and managers could use to (1) determine if driving simulators could help meet training objectives and (2) if so, what kind of simulators to acquire. The end product of this research is a set of task-based criteria that lead to specific simulator characteristics. That is, one should purchase a training simulator based upon what tasks need to be trained. This paper reports on the limited available data on the effectiveness of driving simulators for training, the task clusters various technologies can train, and the decision aids developed for transit agencies that actually have applicability to any potential user of training simulation. The project included a literature review, visits to driving simulator users nationwide, a review of European simulator programs, and the collection of training data and accident data from both users and non-users of driving simulators. Instructors, students, course graduates, and managers were interviewed. The results of the research are presented and a simulator evaluation methodology is proposed.

Younger adults are overinvolved in accidents. Model high school driver education programs were developed in the 1970s in an attempt to reduce this overinvolvement. An evaluation of these programs suggested that they were largely ineffective. Recently, the AAA Foundation for Traffic Safety has developed the first PC-based driver education program (Zero Errors Driving or Driver ZED) using real footage of risky scenarios. The hope is that younger drivers seeing these scenarios will learn to recognize risky situations in the real world before they develop. In an attempt to evaluate the Driver ZED program, the performance of 20 younger drivers is being tested on the University of Massachusetts' driving simulator. Ten of these drivers have been trained with ZED (the trained group) and ten have not seen the program (the untrained group). All 20 drivers must navigate a total of 24 scenarios that have been programmed on the driving simulator. Measures of driving performance were developed which can be used to discriminate between risky and nonrisky drivers. A preliminary evaluation of the performance of the trained and untrained subjects indicates that the trained subjects are performing more cautiously than the untrained subjects in some, but not all, scenarios (e.g., the trained subjects brake sooner when approaching a pedestrian crossing).

Four broad approaches to decreasing the number of cargo tank rollovers were evaluated: driver training, electronic stability aids, improvements in design of the vehicle itself, and highway design. A study of rollover crash statistics confirmed many expectations, but a few of the factors were not as strong as might have been expected. The portion of rollovers that occur on freeways is 15% to 20%. A driver error of one kind or another (e.g., decision or performance error) figures in about three-fourths of cargo tank rollovers. Inattention and distraction account for about 15%. Evasive maneuvers were a factor in 5% to 10% of rollovers. Drivers must be trained to appreciate the diverse causes for rollovers and to anticipate the situations that lead to them. Adherence to viable work and rest schedules is crucial. Electronic stability aids automatically slow the truck when it rounds a curve too fast. They can be remarkably effective in preventing this scenario. However, crash statistics and anecdotal accounts consistently show many other factors that can lead to rollovers. Significant reductions in rollover rates can be achieved with modest changes in vehicle stability. Cargo tank trailers of improved stability are currently available for some cargoes. When mountainous terrain or other factors dictate highway designs that can contribute to rollovers, drivers need to be made aware through signage or dispatch instructions. A comprehensive benefit–cost analysis, conducted from a societal point of view during a 20-year window, projected that the improvements will be cost beneficial.