Macroscopic changes in the urban environment and in the built transportation infrastructure, as well as changes in household demographics and socio-economics, can lead to spatio-temporal variations in household travel patterns and therefore regional travel demand. Dynamics in travel behavior may also simply arise from the randomness associated with values, perceptions, attitudes, needs, preferences and decision-making process of the individual travelers. Most urban travel behavior models and analysis seek to explain variations in travel behavior in terms of characteristics of the individuals and their environment. Spatial extents and temporal variation in an individual’s travel pattern may represent a measure of the individual’s spatial appetite for activity and the variability-seeking nature on his/her travel behavior. The objective of this dissertation effort is to develop a methodology to predict activity participation using revealed spatial extents and temporal variability as variables that represent the spatial appetite and variability-seeking nature associated with individual household. Activity participation is defined as a set of activities in which an individual or household takes part, to satisfy the sustenance, maintenance and discretionary needs of the household. To accomplish the goals of the dissertation, longitudinal travel data collected from the Commute Atlanta Study are used. The raw Global Positioning Systems (GPS) data are processed to summarize trip data by household travel day and individual travel day data. A methodology was developed to automatically identify the activity at the end of each trip. Methods were then developed to estimate travel behavior variability that can represent the variability-seeking nature of the individual. Existing methods to estimate activity space were reviewed and a new Modified Kernel Density area method was developed to address issues with current methods. Finally activity participation models using structural equation modeling methods were developed and the effects of the variability-seeking nature and spatial extent of activities were applied to the models. The variability-seeking nature was presented in the activity participation model as a latent variable with coefficient of variation of trips and distance as indicator variables. The dissertation research found that inclusion of activity space variables can improve the activity participation modeling process to better explain travel behavior.

Les déplacements de personnes et de marchandises sont responsables d’une part importante des impacts environnementaux à l’échelle de la ville. Cette thèse part des hypothèses que, les impacts environnementaux des phases situées en amont ou en aval de la phase d’usage ne sont pas négligeables et que l’étude différenciée des besoins de mobilité peut apporter un éclairage nouveau aux méthodes traditionnelles d’évaluation. La façon de répondre à chaque besoin induit, a priori, des impacts environnementaux différents. Ces derniers sont calculés à l’aide de la méthode d’Analyse de Cycle de Vie. La méthodologie a été développée sur un cas d’application : les émissions de GES de l’Agglomération de Saint-Etienne afin de se confronter à la réalité du terrain et de se procurer des données spécifiques. Ce cas d’application a donné une première série de résultats, confirmant à la fois la part non négligeable d’impacts hors phase d’usage et la disparité des émissions pour une personne.kilomètre selon le besoin de déplacement. Ensuite, une proposition de typologies urbaines a été faite en réalisant une Analyse en Composantes Principales, couplée à une Classification Ascendante Hiérarchique, sur les grandes aires urbaines de France, sur des variables influentes pour la mobilité. Trois classes d’aires urbaines ont été étudiées. La méthodologie a été appliquée aux données des Enquêtes Ménages Déplacements de Bordeaux, Toulon et Valenciennes. Des disparités sont observées pour certains motifs et il est possible de calculer des marges de progression pour chaque agglomération. Enfin, l’utilisation de l’ACV, multicritère a également révélé la nécessité de travailler avec plusieurs indicateurs. / Transportation of goods and people is known as a key environmental impact contributor in a city. However, traditional impact assessment methodologies don’t consider environmental issues all over the life cycle. The cause of a trip is also rarely taken into account in an environmental assessment. This thesis aims at proposing a new impact assessment methodology that considers the influence of indirect emissions (using Life Cycle Assessment (LCA) methodology) and causes of trips on the environmental footprint. Two main hypotheses are settled at the beginning: the environmental impacts of mobility do not only occur during the use phase (the trip by itself) but all over its life cycle. Also the trip purposes may change the choice of way to travel and consequently affect environmental impacts. The global Functional Unit (FU) is “transport people and good in a French urban area during a year” and is divided in intermediate FU referring to each trip purposes, such as “allow commuter trips in a French urban area during a year”. The methodology was developed, tested and validated through a real case: the agglomeration of Saint-Etienne (France). During this phase, mobility territorial actors were met and real data provided that allows facing methodology to possible field limitations. This application case shows first results. First, it confirms importance to consider impacts all over the life cycle (17% of GHG emissions occur out of the use phase). Second, it shows significant differences of emissions by person.kilometer between trip purposes categories. Then a typology of urban area was proposed based on literature about travel determinants. Principal Component Analysis used on French big urban areas (more than 250 000 inhabitants) identified 5 classes. The methodology was applied to three of them (Bordeaux, Toulon and Valenciennes) using data from a standardized survey on households and mobility in order to compare them. It shows disparities on global results partly due to total annual travel distances. The trip purposes differentiation also shows important differences between urban areas. For each cause trip category and emission reduction potentials were calculated for each urban area by extrapolating differences for a pkm to the total annual distance for the trip. This can provide helpful information to deciders. Finally, the need to provide a multicriteria evaluation that is provided by LCA was committed. However methodological improvement would be necessary to better reflect local and regional impacts in LCA.

Traflic safety and accessibility have been two important subjects in transportation research. On the one hand traffic crashes bring about high societal costs and serious health risks for urban road users. The cost oftraffic crashes is estimated to be 17 billion euros per year only in the Netherlands while over 600 people were killed in traffic, of whom 229 were cyclists and 195 were car users [l, 2]. Accessibility, on the other band, is regarded as one of the indicators of the quality of the transport system serving the public. There is comprehensive literature investigating the relationship between traffic crashes and factors associated with traffic, roadway design, built environment, and human factors. Similarly, several studies assessed and evaluated accessibility levels of individuals, communities, and regions by utilizing the aforementioned. factors. Nevertheless, there is a scarcity ofliterature investigating the relationships between accessibility and traffic safety. This is especially surprising considering that both subjects are associated with a similar set of factors, including land use and transport systems, as weil as individual and temporal factors [3-7]. The relationships between accessibility and traffic safety can be an adverse one; for example, improved accessibility by increasing the travel speeds (i.e., declining travel time) intensifies the crash risks which also deteriorates equity. Furthermore, levels ofboth accessibility and traffic safety are not homogeneous throughout urban areas and among different population groups. Based on the literature, it is obvious that accessibility is associated with economic equity [8]. lt is revealed that accessibility of lower-income groups is substantially worse than the higher-income groups as these groups have less mobility [9]. Previous studies also showed. that lower-income groups usually suffer from traffic safety problems more than other socio-economic groups [10-12]. Therefore, this research aims to address the aforementioned gap in the literature in understanding the relationships between accessibility levels and traffic safety with a focus on social equity perspecti.ves. For this purpose, a Gravity model and risk exposure evaluation approaches are utilix.ed to analyze traffic safety and accessibility to jobs by bicycle via extending the traditional definition of accessibility based on only travel time or proximity to a location.

No description available.

Contexte: La réduction de l'utilisation de l’automobile a déjà été identifiée comme étant une intervention populationnelle pouvant promouvoir l’activité physique. Les recherches portant sur la relation entre l’utilisation de l’automobile et l’activité physique utilisent typiquement une variable catégorielle pour décrire le mode de transport, tiennent peu compte du motif de déplacement et n’explorent pas la variation de l’association selon le lieu de résidence. Cette étude utilise à la fois l'activité physique auto-rapportée et objectivement mesurée pour tester l'association entre l’utilisation de l’automobile et l’activité physique totale, tout en tenant compte de ces limitations antérieures. Méthode: Les données de 780 participants provenant de la branche montréalaise de l’étude INTERACT (2018-2019) ont été analysées. Des modèles de régression linéaire ont été construits afin d’examiner la relation entre l’utilisation de l’automobile et l'activité physique modérée-vigoureuse (APMV) totale (auto-rapportée avec transformation logarithmique et objectivement mesurée). Ensuite, le motif de déplacement a été inclus dans ces modèles. Enfin, une régression pondérée géographiquement a permis d’explorer la variation spatiale de l'association entre l’usage de l’automobile et l’APMV. Résultats: La proportion des déplacements effectués en automobile est associée négativement avec l'APMV totale auto-rapportée (coefficient : -0,009, intervalle de confiance à 95% : -0,012 à -0,006) et objectivement mesurée (Coefficient : -0,29 minute par jour, intervalle de confiance à 95% : -0,55 à -0,03). La régression pondérée géographiquement indique une faible variation spatiale de l'association entre l'utilisation de l’automobile et l’APMV totale auto-rapportée. Le nombre de déplacements pour un motif tel que commerce et services est associé à l'APMV totale. Conclusion: Cette étude transversale a démontré une association négative entre l’utilisation de l’automobile et l’APMV totale sur le territoire montréalais. Certains motifs de déplacement sont associés à l’activité physique totale. / Background: Car usage reduction has previously been pointed out to be a population-based intervention promoting physical activity. Previous literature on the car usage-physical activity relation typically uses categorical variables for transport modes, rarely accounts for trip purpose and does not explore the influence of home location on this association. This study uses both self-reported and objectively measured physical activity to test the association between car usage and total daily physical activity, while accounting for previous limitations. Methods: INTERACT data collected in 2018-2019 among 780 participants from the Montreal metropolitan area site were analysed. Linear regression models of self-reported (log-transformed) and objectively measured total moderate-vigorous physical activity (MVPA) were built separately, as a function of car usage. Trip purpose was then included in these models. Finally, a geographically weighted regression (GWR) model was built to explore the spatial variation of the car usage-MVPA association. Results: The proportion of trips made by car showed a negative association with both self-reported (coefficient: -0.009, 95% CI [-0.012, -0.006]) and objectively measured (coefficient: -0.29 minutes per day, 95% CI [-0.55, -0.03]) total MVPA. GWR showed little spatial variation in the car usage-total self-reported MVPA. The number of trips toward certain purposes (e.g. “shops and services”) is associated with total MVPA. Conclusion: This cross-sectional study showed a negative association between car usage and total MVPA in the Montreal metropolitan area. Some trip purpose is associated with total physical activity.