|
Abstract
|
The experiment described herein was conducted to develop a method of predicting driver ability based on objective measures of driving performance.
Fifteen subjects drove an instrumented vehicle in live traffic on two-lane and multi-lane road sections in and around the Toronto area. Eight of the subject drivers had at least five years driving experience and seven of them had less than two thousand miles driving experience.
On selected portions of the road course, subjects were instructed to maintain certain speeds or lanes. During these periods raw data were collected on vehicle velocity, lane position, steering wheel position, and accelerator pedal position. Using off-line computer programs the data were transformed into a number of descriptive statistics which were then analyzed using univariate and multivariate statistical techniques.
Results indicated that even though univariate analyses were generally unsuccessful in differentiating between the groups of experienced and inexperienced drivers, successful discrimination was achieved with combinations of variables.
Results suggest that it could be possible to employ on-line monitoring devices to determine whether a driver is capable of a minimum level of driving performance. Implications for the use of such a device in driver licensing and education are discussed.
Language: en |