自动驾驶技术给出行领域带来的机遇是否会顺利显现，很大程度上取决于公众对自动驾驶出行服务的使用意愿和使用模式。部分城市推出的基于自动驾驶（有安全员）的试乘和运营服务为公众了解新技术提供了契机，也为探索未来的出行变化提供了新视角。因此，本文致力于解析使用意愿的形成机理，并从自动驾驶使用模式的分类情况、车内时间利用和旅行时间变化的影响因素多个角度出发掌握使用模式的规律和特征，从而为潜在的出行行为变化提供见解。 首先，针对免费试乘体验过自动驾驶小巴（有安全员）的试乘用户，本文构建了包含信任和体验满意度的扩展计划行为理论模型，并以社会人口学特征作为调节变量，从而研究试乘用户对自动驾驶汽车使用意愿的形成机理。采用结构方程模型分析发现体验满意度能通过信任、使用态度和感知行为控制来间接影响使用意愿，但体验满意度对感知行为控制的影响会随性别和受教育程度而变化。 进一步，本文探究了经验用户在多次乘坐公开、收费的自动驾驶出租车（有安全员）中的使用特征。为研究经验用户持续使用该出行服务的影响因素，本文构建了包含使用态度、服务满意度、平等安全担忧、心理拥有感的拓展技术接受模型，并分析了情景因素的调节效应。结构方程模型分析发现使用态度是影响力最大的因素，安全员频繁接管车辆正向地调节了平等安全担忧与使用态度之间的关系。 接着在针对使用模式的分类研究中，本文以当前自动驾驶出租车市场显在用户和潜在用户为研究对象，构建多维度的使用行为外显变量来考虑自动驾驶带来的广泛影响，进而运用潜在类分析方法识别出了三类旅行者，掌握了不同类旅行者未来使用模式的特征，明确了影响类别隶属度的的关键因素，检验了不同类旅行者对自动驾驶态度的差异，为理解用户需求、出行服务优化提供支撑。 最后，本文建立了自动驾驶背景下车内时间利用偏好和通勤时间变化影响因素概念模型，分别以自驾通勤者和公交通勤者为研究对象，构建考虑该概念模型的排序Probit和随机参数有序Logit模型来对两类通勤者的未来车内时间利用偏好和通勤时间变化进行分析。结果解析了车内时间利用偏好的影响因素，验证了当下和未来的车内时间利用类型将具有连续性，同时揭示了自动驾驶从旅行时间的感知与期待、自动驾驶的接触和感知情况、未来车内时间利用偏好三个方面影响通勤时间增加程度的潜在机理。
Whether the opportunities brought about by autonomous driving will emerge in the field of mobility largerly depends on the public‘s usage intention and usage patterns for autonomous driving mobility services. Some cities have launched test ride projects and operational services based on autonomous driving, which provides opportunities for the public to understand autonomous driving, and also provides a new perspective for exploring future travel behavior changes. Therefore, this study aims to analyze the formation mechanism of usage intention, and grasp the rules and characteristics of usage patterns from multiple perspectives, namely, the classification of usage patterns of autonomous driving mobility services, the influencing factors of travel time utilization and travel time changes. The results will provide insights into potential changes in travel behavior. First of all, for the test ride users who have experienced the autonomous minibus (with safety steward) for free, this study constructs an extended planned behavior theory model including trust and experience satisfaction, and takes socio-demographic attributes as moderators, so as to study the formation mechanism of test riders‘ usage intention of autonomous vehicle. Based on the results of structural equation model analysis, it is found that experience satisfaction can indirectly affect usage intention through trust, attitude, and perceived behavioral control, but the impact of experience satisfaction on perceived behavioral control varies with gender and education level. Further, this study explores the usage characteristics of experienced users who have repeatedly taken public and paid autonomous taxis. At the same time, in order to study the influencing factors of experienced users‘ continuous use intention of autonomous taxis, this study constructs an extended technology acceptance model including attitude, service satisfaction, concern for the equality of safety, and psychological ownership, and analyzes the moderating effect of the contextual factor (i.e., safety steward takeover frequency). Structural equation model analysis shows that attitude is the most influential factor, and the frequent takeovers by safety steward positively moderates the relationship between concern for the equality of safety and attitude. Then, in the classification study of usage patterns, this study puts forward multi-dimensional manifest variables of usage behavior to consider the extensive impact of autonomous driving, and uses the latent class analysis method to identify three types of travelers among existing and potential users in the current autonomous taxi market. Furthermore, this study grasps the characteristics of future usage patterns of different types of travelers, clarifies the key factors affecting class membership, and tests the differences in attitudes towards autonomous driving among different types of travelers, providing support for understanding user needs and optimizing mobility service. Finally, this study establishes a conceptual model of factors affecting travel time utilization preference and commuting time changes under the background of autonomous driving. Taking private car commuters (as driver) and public transport commuters as research objects, this study constructs rank-ordered probit models and random parameter ordered logit models considering the conceptual model to analyze the changes of time utilization preference and commuting time of the two types of commuters. The results analyze the influencing factors of travel time utilization preference, and verify the continuity of the current and future travel time utilization types. Also, the results reveal the potential mechanism by which autonomous driving affects the increase level of commuting time from three aspects: perception and expectation of travel time, exposure and perception of autonomous driving, and future travel time utilization preference.
