Sensitivity analysis of driver's behavior and psychophysical conditions

To reduce traffic accidents, an accurately estimated model is needed to capture the true relationships between the injury severity and risk factors. This study aims to propose a robust procedure to address the biases in police-reported accident data and subsequently to conduct sensitivity analyzes i...

Full description

Bibliographic Details
Authors: García Herrero, Susana, Gutiérrez, José Manuel, Herrera, Sixto, Azimian, Amin, Mariscal Saldaña, Miguel Ángel
Format: article
Status:Versión aceptada para publicación
Publication Date:2020
Country:España
Institution:Universidad de Burgos (UBU)
Repository:Repositorio Institucional de la Universidad de Burgos (RIUBU)
OAI Identifier:oai:riubu.ubu.es:10259/9037
Online Access:http://hdl.handle.net/10259/9037
Access Level:Open access
Keyword:Traffic accident
Drugs
Alcohol
Speed
Distraction
Human error
Bayesian network
Bias identification
Transportes
Salud
Transportation
Health
Description
Summary:To reduce traffic accidents, an accurately estimated model is needed to capture the true relationships between the injury severity and risk factors. This study aims to propose a robust procedure to address the biases in police-reported accident data and subsequently to conduct sensitivity analyzes in order to estimate the variations in injury severity and distraction probability based on drivers’ behaviors/characteristics and psychophysical conditions. The results show that: (i) the excess speed will likely increase the probability of serious/fatal injury for drivers of all age groups by 10%; (ii) distraction and driver’ errors will likely increase the probability of serious/fatal injury in all drivers driving at a proper speed up to 1.5%; (iii) alcohol and drug consumption can significantly increase the probability of being distracted and making errors by 28.5% and 33.5% respectively; (iv) Alcohol consumption reduces the probability of driving at an appropriate speed in drivers under 25 by 40%. However, the results for drugs consumption are not as significant as the ones for alcohol consumption.