【摘要】 随着交通系统自组织运行与自主化服务能力的快速发展，自主式交通系统（Autonomous Transportation Systems, ATS）成为智能交通系统未来的重要发展方向，构建ATS体系架构是规划与建设新一代智能交通系统的基础。面向典型交通场景，首先要分析交通系统的用户需求，本文重点解决基于ATS需求体系的典型交通场景用户需求分析问题，具体研究场景关键词与需求文本库的匹配方法。针对ATS的场景需求分析问题，首先提出基于分层的场景描述方法，对不同抽象层次的交通场景进行分层分解，并根据活动理论解析各抽象层次场景的属性要素。其次，根据属性要素特点提出基于短文本匹配的双层匹配模型，在双层模型中应用了TF-IDF和LSI两种相似度匹配算法。最后，在一体化出行服务典型场景中进行方法应用与对比，判断匹配模型和算法的实际效果。结果表明，采用TF-IDF和LSI算法协同作用的双层匹配模型取得了较好的匹配效果，匹配度指标达85%，并进一步通过灵敏度分析验证了双层匹配方法的适用性。更多还原

【Abstract】 Owing to the rapid development of the self-organized operation and autonomous service capability of transportation systems, Autonomous Transportation Systems(ATS) have been investigated extensively. To establish the architecture of the system, user requirements under typical traffic scenarios must first be analyzed. This paper focuses on user requirement analysis under typical traffic scenarios involving ATS. The key is to develop a method that can match the keywords of scenarios with information regarding user needs from a specified ATS database. First, a hierarchical description method is proposed to decompose the traffic scenarios into different abstraction levels. Subsequently, the scenario attributes at different levels are analyzed based on activity theory. Second,based on the scenario attributes, a two-level matching model based on short-text matching is proposed. Two similarity matching algorithms, i.e., TF-IDF and LSI, are applied in the two-level matching model. Finally, the proposed method is applied to a scenario involving Mobility as a Service, and the performances of the matching model and algorithm are evaluated. Numerical results show that the two-level matching model with TF-IDF and LSI algorithms achieves good matching performance, with a matching index of 85%. Additionally, the applicability of the two-level matching method is verified via sensitivity analysis.更多还原