Public opinion on autonomous vehicles (AVs) is an important topic as AVs are expected to change the transportation system dramatically. Although the topic has been discussed extensively by many researchers and experts, it is beneficial to complement current studies by explaining the AV-related sentiment variations leveraging the unique social media features.

Different from traditional survey-based studies, this study relies on tweets to understand the sentiments on AVs. A comprehensive model framework is proposed to categorize sentiments, recognize critical dates with sentiment changes, and explain sentiment variations. The results indicate that the general sentiment is positive towards AVs, but social media users have sentiment biases towards different AV terms. Significant sentiment changes are often linked with major social events related to AVs, and the general public is more sensitive to social events than the most active users.

A wide range of policy insights are discussed based on the results of the analyses, including policies related to safety, pricing, unemployment, and AV adoption rate. Possible challenges and the corresponding strategies are also discussed. These insights will be helpful for the public agencies, automobile manufacturers, and technology companies in gaining a better understanding of AV adoption, and in preparing for the future of transportation systems.

关于自动驾驶汽车 (AV) 的公众舆论是一个重要话题，因为预计 AV 将极大地改变交通系统。尽管该主题已被许多研究人员和专家广泛讨论，但通过利用独特的社交媒体功能解释与 AV 相关的情绪变化来补充当前的研究是有益的。

与传统的基于调查的研究不同，这项研究依赖于推文来了解对 AV 的情绪。提出了一个综合模型框架来对情绪进行分类，识别具有情绪变化的关键日期，并解释情绪变化。结果表明，总体情绪对 AV 是积极的，但社交媒体用户对不同的 AV 术语有情绪偏见。显着的情绪变化通常与与 AV 相关的重大社会事件有关，普通大众比最活跃的用户对社会事件更为敏感。

根据分析结果讨论了广泛的政策见解，包括与安全、定价、失业和自动驾驶汽车采用率相关的政策。还讨论了可能的挑战和相应的策略。这些见解将有助于公共机构、汽车制造商和技术公司更好地了解 AV 的采用，并为交通系统的未来做好准备。