The aim of this paper is to investigate the safety of cyclists while interacting with autonomous vehicles (AVs) in an uncontrolled T-intersection using PTV Vissim microsimulation combined with the Surrogate Safety Assessment Model (SSAM). For the microsimulation, three cycling behaviours were constructed (cautious, normal, and aggressive), and two AV behaviours (cautious and normal) were adopted. 87 h of microsimulation resulted in 230 conflicts. The type and nature of these conflicts, as well as two temporal indicators and four speed-related indicators were analysed. Nearly 96% of the conflicts were rear-end collisions, and almost 70% were conflicts where the cyclist was the first to arrive at the conflict point. Interestingly, the least involved cycling behaviour in conflicts was aggressive and the most involved was cautious. A significant relationship was found between severity clusters of temporal indicators and cycling behaviour. It is also concluded that an important future research direction would be the construction of a two-dimensional conflict diagram with severity thresholds specifically for bicycle-vehicle interactions.

中文翻译：

---

骑车人与自动驾驶车辆互动的安全性：微观模拟和 SSAM 分析相结合

本文的目的是使用 PTV Vissim 微观模拟与替代安全评估模型 (SSAM) 相结合，研究骑车者在不受控制的丁字路口与自动驾驶车辆 (AV) 交互时的安全性。对于微观模拟，构建了三种骑行行为（谨慎、正常和激进），并采用了两种 AV 行为（谨慎和正常）。 87 小时的微观模拟导致 230 起冲突。分析了这些冲突的类型和性质，以及两个时间指标和四个与速度相关的指标。近96%的冲突是追尾事故，近70%是骑车人首先到达冲突点的冲突。有趣的是，冲突中参与最少的骑行行为是攻击性的，而参与最多的则是谨慎的。时间指标的严重程度集群与骑行行为之间发现了显着的关系。还得出的结论是，未来一个重要的研究方向将是构建专门针对自行车与车辆交互的具有严重性阈值的二维冲突图。