In conjunction with a large law-enforcement agency, MFES is involved in some exciting research into driver attitudes toward driving and accidents, physical performance of the driving task, and accident involvement. The ultimate goal is to reduce the number of accidents in which the department is involved. With complete test data on the performance characteristics of hundreds of drivers as they piloted DAQ-equipped vehicles around a test-course, and the opportunity to correlate this performance to individual drivers' accident involvement may provide the most comprehensive bank of information commonly available on driver abilities and how they relate to accident inlvement. It is hoped that this research will reveal new clues about who gets into accidents, and perhaps why. This information should help guide modifications to the training program to reduce accidents.
Two vehicles have been used: a 1996 Chevrolet Lumina (front-wheel-drive), with the 3.1 liter V-6 engine, weighing approximately 3400 pounds, and a 1991 Chevrolet Caprice (rear-wheel-drive), equipped with a 5.7 liter V-8, weighing approximately 4100 pounds. Each car had a PC-based data acquisition package which recorded the steering wheel position, brake line pressure, throttle position, lateral acceleration, and longitudinal acceleration at 50 Hz during a "test-run" as the vehicle was piloted around a cone-demarcated course.
The course was laid out on flat pavement at a driver training facility. Beginning from a stop, the drivers travelled straight ahead for approximately 250 feet, then turned right into a 100 foot radius circle. After 1.5 revolutions, they turned right again, travelling in the same path they had earlier traversed. After a straight path of about 300 feet, there was a decreasing radius turn, culminating in a sharp right turn, a sharp left turn, and another sharp right turn. This was followed by a slalom portion, so designed as to require approximately 180 degrees of steering wheel excursion to negotiate each gate. The run ended in a straight approach to a brake chute.
With the number of tests (hundreds), the database is very large, and work to condense this information into useful fragments is ongoing.
One portion of the work included converting the steering output to degrees of steering wheel rotation, from which steering wheel rates could be calculated. The following picture shows a fraction of the steering output by one driver (pulled at random from the database), along with the steering rates generated over 1 tenth and 3 tenths of a second.
Publication of this work will be ongoing, as is appropriate. To date, no conclusions have been drawn about accident involvement, but there have been some surprises concerning driver attitudes and capabilities.
POSTSCRIPT: Sadly, this project was more or less cancelled by someone higher up the foodchain than those of us involved in it, and no crash statistics were gathered for comparison to behavior. An SAE paper did come of it, though, 2000-01-0179
Copyright 1999 by Wade Bartlett
Mechanical Forensics Engineering Services, LLC.
This page last modified 04-AUG-2020