This article deals with the application of fuzzy techniques to the CREAM model of Hollnagel, already recognized in the scientific community in applications on the human component in complex systems. Among those most applied, the CREAM has the advantage of being extremely flexible and of being able to receive the support of opportune analytical instruments for the calculation of human reliability. In this article, the uncertainty of the environment within which driving takes place has been represented through fuzzy techniques. In particular, the variables that describe the state of the con-text perceived by the driver have been modeled by means of appropriate membership functions, char-acterized by lexical-type domains and nonlinear dependencies. The results obtained from the consequent process of defuzzification have been further trans-formed into intervals in order to conserve evidence of the uncertainty also subsequent to the analy-sis, up to determination of the reliability of the human component. The goals achieved, evidenced in a simple application, can be used in accident analysis but, above all, at the previsional stage, that is, the designing of new or existing roads, when we wish to deepen the presumed criticalities that haven’t emerged from the rigorous application of the current standard.
THE APPLICATION OF DRIVERS’ COGNITIVE RELIABILITYUSING FUZZY TECHNIQUES
PELLEGRINO, Orazio
2012-01-01
Abstract
This article deals with the application of fuzzy techniques to the CREAM model of Hollnagel, already recognized in the scientific community in applications on the human component in complex systems. Among those most applied, the CREAM has the advantage of being extremely flexible and of being able to receive the support of opportune analytical instruments for the calculation of human reliability. In this article, the uncertainty of the environment within which driving takes place has been represented through fuzzy techniques. In particular, the variables that describe the state of the con-text perceived by the driver have been modeled by means of appropriate membership functions, char-acterized by lexical-type domains and nonlinear dependencies. The results obtained from the consequent process of defuzzification have been further trans-formed into intervals in order to conserve evidence of the uncertainty also subsequent to the analy-sis, up to determination of the reliability of the human component. The goals achieved, evidenced in a simple application, can be used in accident analysis but, above all, at the previsional stage, that is, the designing of new or existing roads, when we wish to deepen the presumed criticalities that haven’t emerged from the rigorous application of the current standard.Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
