Citation

Parker RA. J. Saf. Res. 1977; 9(2): 59-67.

Copyright

(Copyright © 1977, U.S. National Safety Council, Publisher Elsevier Publishing)

DOI

unavailable

PMID

unavailable

Abstract

The effectiveness of motor vehicle safety inspections can be significantly improved if vehicles are examined with a frequency based on the likelihood of component failure. A probabilistic model of motor vehicle degradation is developed using a set of Weibull distributions coupled with mixing distributions that account for the component criticality and vehicle make composition of the subject vehicle population. Using available data from existing vehicle inspection programs, the parameters of the model can be estimated. Techniques for this estimation and sample parameter values are presented. A computer algorithm is described that allows the model to be used as a means for selecting vehicles for inspection. A test of the model is discussed, where in vehicles were selected by the model and compared with a similar sample drawn purely at random. Statistical comparison of the two samples indicates an improvement in the overall inspection effectiveness of 37%.