ABSTRACT
In 2002, the University of Iowa conducted a study of driver responses to a simulated rear tread belt detachment using the National Advanced Driving Simulator (NADS). The researchers evaluated the ability of test subjects to maintain control of simulated vehicles with this simulated tire failure. They also evaluated the hypothesis that a vehicle's linear range understeer gradient would affect that ability. The vast majority of subjects that “drove” the simulated vehicle that was closest to a real-world vehicle had no difficulty maintaining control using steering and braking, a result consistent with real-world findings in other studies. Contrary to real world findings in other studies, the majority of Ranney NADS subjects failed to realize that a tire failure had occurred and they displayed no appropriate response to the event itself. Further, those subjects that did lose control did so as a result of extremely small steering commands inconsistent with the recorded remaining cornering capacities of real vehicles. In order to resolve those apparent conflicts among studies, the available data underlying the Ranney NADS study were obtained with the cooperation of the National Highway Traffic Safety Administration including drivers’ steering, braking, and acceleration inputs before, during and after the simulated tread belt detachment. In this paper, these data as well as available information regarding the NADS protocols and the NADS computer modeling, including vehicle and tire parameters, were analyzed to identify sources of the conflicts when compared to real world studies. This independent analysis concludes that the conflicts arise because the Ranney NADS study methodology is significantly flawed and because its underlying data do not support the conclusions stated by its authors and should not be used as a basis for a human factors study.
