Publicação: Safety performance models for urban intersections in Brazil
Carregando...
Data
Orientador
Coorientador
Pós-graduação
Curso de graduação
Título da Revista
ISSN da Revista
Título de Volume
Editor
Elsevier B.V.
Tipo
Artigo
Direito de acesso
Acesso restrito
Resumo
This paper presents a modeling effort for developing safety performance models (SPM) for urban intersections for three major Brazilian cities. The proposed methodology for calibrating SPM has been divided into the following steps: defining the safety study objective, choosing predictive variables and sample size, data acquisition, defining model expression and model parameters and model evaluation. Among the predictive variables explored in the calibration phase were exposure variables (AADT), number of lanes, number of approaches and central median status. SPMs were obtained for three cities: Fortaleza, Belo Horizonte and Brasilia. The SPM developed for signalized intersections in Fortaleza and Belo Horizonte had the same structure and the most significant independent variables, which were AADT entering the intersection and number of lanes, and in addition, the coefficient of the best models were in the same range of values. For Brasilia, because of the sample size, the signalized and unsignalized intersections were grouped, and the AADT was split in minor and major approaches, which were the most significant variables. This paper also evaluated SPM transferability to other jurisdiction. The SPM for signalized intersections from Fortaleza and Belo Horizonte have been recalibrated (in terms of the COx) to the city of Porto Alegre. The models were adjusted following the Highway Safety Manual (HSM) calibration procedure and yielded C-x of 0.65 and 2.06 for Fortaleza and Belo Horizonte SPM respectively. This paper showed the experience and future challenges toward the initiatives on development of SPMs in Brazil, that can serve as a guide for other countries that are in the same stage in this subject. (C) 2014 Elsevier Ltd. All rights reserved.
Descrição
Palavras-chave
Safety performance models, Road safety, Generalized linear models, Observational road safety studies, Generalized Estimating Equations
Idioma
Inglês
Como citar
Accident Analysis And Prevention. Oxford: Pergamon-elsevier Science Ltd, v. 70, p. 258-266, 2014.
