Understanding and Modeling Non-motorized Travel Behavior

可持续城市规划旨在通过降低汽车作为首选出行方式的重要性来提高生活质量。实现这一目标的策略包括步行社区、混合使用环境、以交通为导向的发展、密集化、完整的街道和多式联运系统。然而，要预测这些策略的影响，需要将所有出行方式纳入预测模型，而不仅仅是汽车和公共交通。鉴于加拿大的城市规划现在侧重于可持续性，我们有必要进一步了解非机动车出行行为——步行和骑自行车——并开发可用于评估上述策略的预测模型。为了进一步发展这种模型，本研究有两个相互交织的目标：(1)加强对加拿大非机动化旅行行为的理解；(2)解决模型从一个时空背景到另一个时空背景的可转移性问题。为实现这些目标，设计了五个目标：(1)分别为工作和非工作旅行制定“国家级”模式选择模型，其中包括步行和骑自行车；(2)建立“国家级”出行模式选择模型，将工作出行和非工作出行两个时间段的数据分别纳入步行和骑自行车；(3)建立“国家级”出行生成模型，捕捉不同出行方式的日均出行数量之间的相互作用；(4)建立“国家级”出行生成模型，将两个时间段的数据汇集在一起，捕捉不同模式下每日出行数量之间的相互作用；(5)分别建立了步行和骑自行车的路径选择模型。***关于推进知识，拟议的研究将揭示控制非机动旅行的因素（例如，天气条件，建筑环境，人的特征）在全国范围内导致相同的行为结果的程度。这项研究还将揭示这些因素的影响是否会随着时间的推移而改变，从而改变行为结果。就实际成果而言，通过本研究开发的政策敏感模型可以很容易地传播到加拿大各地的城市规划机构，用于预测可持续城市规划战略引起的步行和骑自行车需求的变化。这些模型对目前没有这种预测工具的机构将特别有用。总之，本文提出的研究不仅促进了知识的发展，而且为开发具有成本效益且行为合理的旅行需求预测模型提供了创新的解决方案

Sustainable urban planning seeks to enhance quality of life by deemphasizing cars as a preferred means of travel. The strategies for achieving this include walkable neighborhoods, mixed-use environments, transit-oriented development, densification, complete streets, and multi-modal transportation systems. However, predicting the impacts from such strategies requires forecasting models that incorporate all modes of travel, not just cars and transit. Given that urban planning in Canada now focuses on sustainability, there is a timely need to further our understanding of non-motorized travel behavior – walking and bicycling – and develop forecasting models that can be used to evaluate the strategies listed above.*** To further such model development, the proposed research has two intertwined goals: (1) to enhance the understanding of non-motorized travel behavior in Canada, and (2) to address the issue of model transferability from one spatio-temporal context to another. Five objectives are designed to achieve these goals: (1) develop “national-level” mode choice models for work and non-work trips, respectively, that incorporate walking and bicycling; (2) develop “national-level” mode choice models pooling data from two time periods for work and non-work trips, respectively, that incorporate walking and bicycling; (3) develop a “national-level” trip generation model that captures interactions among the number of daily trips by mode; (4) develop a “national-level” trip generation model pooling data from two time periods that captures interactions among the number of daily trips by mode; and (5) develop route choice models for walking and bicycling, respectively.*** With respect to advancing knowledge, the proposed research will reveal the extent to which factors governing non-motorized travel (e.g., weather conditions, built environments, characteristics of people) lead to the same behavioral outcomes across the country. The research will also reveal whether the impact of such factors changes over time, therefore altering behavioral outcomes. In terms of practical outcomes, the policy-sensitive models developed through this research can be easily disseminated to urban planning agencies across Canada for forecasting changes in walking and cycling demand arising from sustainable urban planning strategies. The models will be especially useful to agencies without such forecasting tools at the present time. To summarize, the proposed research not only advances knowledge, it provides an innovative solution to developing cost-effective travel demand forecasting models that are behaviorally sound.**