Distâncias de viagens intrazonais: abordagem para estimativa e aplicação a um estudo de caso
Data
2017-11-15
Autores
Título da Revista
ISSN da Revista
Título de Volume
Editor
Biblioteca Digital de Teses e Dissertações da USP
Universidade de São Paulo
Escola de Engenharia de São Carlos
Universidade de São Paulo
Escola de Engenharia de São Carlos
Resumo
Descrição
Este estudo teve como objetivo o desenvolvimento de procedimentos para estimar distâncias de viagens intrazonais com base em dados de pesquisas origem-destino (O/D). A análise conduzida na aplicação apresentada foi baseada em dados de uma pesquisa O/D realizada na cidade de São Carlos, Brasil. A área urbana foi subdividida várias vezes para produzir 471 zonas de análise de tráfego (TAZ) para análise. A análise partiu da identificação de características geométricas das zonas de análise de tráfego (TAZ) que podem influenciar as distâncias das viagens intrazonais. Em seguida, indicadores numéricos dos padrões geométricos selecionados foram comparados com os valores médios das distâncias intrazonais, a fim de procurar por evidências de correlação entre estas variáveis. Duas abordagens analíticas foram exploradas: i) contínua e ii) descontínua. No primeiro caso, todas as viagens foram consideradas em um único conjunto de dados, enquanto que no segundo caso, o conjunto de dados foi dividido em dois subgrupos com características homogêneas. Os conjuntos de dados com maior correlação foram usados para construir modelos globais e estratificados para estimar distâncias intrazonais. Nos modelos globais, viagens a pé, de bicicleta e de automóvel foram analisadas como uma única base de dados. Nos modelos estratificados, os três modos de transporte foram considerados separadamente. Foram ajustados, também, modelos adicionais que permitam estimar as viagens por um modo de transporte específico a partir de valores globais de distâncias intrazonais. Os valores obtidos com estes modelos foram então comparados com as estimativas de outros modelos encontrados na literatura. Os modelos descontínuos aqui obtidos superaram claramente os modelos tradicionais. Isto sugere que podem constituir alternativas para substituir os modelos tradicionais para estimar distâncias médias de viagens intrazonais, pelo menos no caso das cidades médias brasileiras.
This study aimed at the development of procedures for estimating intrazonal trip distances based on data of origin-destination (O/D) surveys. The analysis conducted in the application presented was based on an O/D survey conducted in the city of São Carlos, Brazil. The urban area was subdivided several times to produce 471 traffic analysis zones (TAZ) for analysis. Geometrical characteristics of the zones that can influence the distances of intrazonal trips were then identified. Next, numerical indicators of the selected geometrical patterns were compared with the average values of intrazonal trip distances in order to look for evidences of correlation between these variables. Two analytical approaches were explored: i) continuous and ii) discontinuous. In the first case, all trips were consideredas a single dataset, whereas in the second case the dataset was split in two subsets with homogeneous characteristics. The datasets with higher correlation were used to build global and stratified models for estimating intrazonal trip distances. In the global models, walking, cycling, and auto trips were all considered as part of a common dataset. In the stratified models, the three transport modes were separately taken into account. Additional models, which were meant to estimate trips per mode based on global values of intrazonal trip distances, were also developed. The values obtained with these models were then compared with the estimations of other models found in the literature. The discontinuous models calibrated in this study clearly outperformed the traditional models. This may be an indication that they can be used to replace the traditional models for estimating average intrazonal trip distances, at least in the case of Brazilian medium-sized cities.
This study aimed at the development of procedures for estimating intrazonal trip distances based on data of origin-destination (O/D) surveys. The analysis conducted in the application presented was based on an O/D survey conducted in the city of São Carlos, Brazil. The urban area was subdivided several times to produce 471 traffic analysis zones (TAZ) for analysis. Geometrical characteristics of the zones that can influence the distances of intrazonal trips were then identified. Next, numerical indicators of the selected geometrical patterns were compared with the average values of intrazonal trip distances in order to look for evidences of correlation between these variables. Two analytical approaches were explored: i) continuous and ii) discontinuous. In the first case, all trips were consideredas a single dataset, whereas in the second case the dataset was split in two subsets with homogeneous characteristics. The datasets with higher correlation were used to build global and stratified models for estimating intrazonal trip distances. In the global models, walking, cycling, and auto trips were all considered as part of a common dataset. In the stratified models, the three transport modes were separately taken into account. Additional models, which were meant to estimate trips per mode based on global values of intrazonal trip distances, were also developed. The values obtained with these models were then compared with the estimations of other models found in the literature. The discontinuous models calibrated in this study clearly outperformed the traditional models. This may be an indication that they can be used to replace the traditional models for estimating average intrazonal trip distances, at least in the case of Brazilian medium-sized cities.
Palavras-chave
Transporte em cidades médias, Viagens de bicicleta, Distâncias de viagens intrazonais, Viagens de automóvel, Viagens a pé, Auto trips, Transport in medium-sized cities, Intrazonal trip distances, Cycling trips, Walking trips