Driving simulator dataset on human driven vehicles' gap acceptance behaviour in mixed traffic with automated vehicles

This research is based on data gathered in 2020 at Delft University of Technology. This dataset is from a driving simulator experiment, whose goal was to study human drivers' behaviour in mixed traffic, which has both human-driven vehicles and automated vehicles (AVs).<br>The dataset consists of 95 participants of which 71 (74.7 %) were male and 24 females.<br>The route in the driving simulator consisted of several motorway sections, provincial (regional) road sections, and three priority T-intersections. Each T-intersection consisted of an urban road (the minor road) intersecting with a provincial road (the major road). The defined speed limit was 100 km/h on the motorway, 80 km/h on the provincial roads, and 50 km/h on urban roads. On the motorway, drivers also experienced dynamic speed limit sections. A depiction of the route is attached in this dataset information.<br>The experiment design aimed to separately observe the effects of AVs’ recognizability and their driving style on human driving behavior as well as their combined effects. Each participant drove four scenarios, excluding a familiarization drive. The scenarios differed in two aspects: the recognizability and the driving style of AVs.Two variables primarily varied in the experiment: the driving style of AVs, and their recognizability. The participants were assigned randomly to one of three groups: Defensive AVs, Aggressive AVs, and Mixed AVs. The group determined the driving style of AVs that a participant encountered in the experiment.<br>More details about the experiment set-up itself can be found in our paper published: https://doi.org/10.1016/j.trf.2022.09.018<br><strong>Attached files:</strong>Processed datasetReadMe fileData processing scripts<br><strong>File formats:</strong>Data /.csvData /.xlsxJupyternotebooks /.ipynb

本研究基于代尔夫特理工大学（Delft University of Technology）2020年采集的实验数据。该数据集源自一项驾驶模拟器实验，其研究目标为探究混合交通场景下人类驾驶员的行为规律——此类场景同时涵盖人工驾驶车辆与自动驾驶汽车（Automated Vehicles, AVs）。
本数据集共纳入95名受试者，其中71名（占比74.7%）为男性，剩余24名为女性。
驾驶模拟器的测试路线包含多段高速公路、省道（区域道路）路段，以及3处优先通行T型交叉口。每处T型交叉口均由一条城市支路（次要道路）与一条省道（主要道路）交汇形成。实验设定的限速标准为：高速公路100 km/h，省道80 km/h，城市道路50 km/h。在高速公路路段，驾驶员还会途经动态限速区域。本数据集说明中附带了该测试路线的示意图。
本实验设计旨在分别探究自动驾驶汽车的可识别性、驾驶风格对人类驾驶行为的单独影响，以及二者的联合影响。每名受试者需完成4组测试场景，不包含熟悉驾驶环节。实验场景在两个维度上存在差异：自动驾驶汽车的可识别性，以及其驾驶风格。实验中主要调控两个变量：自动驾驶汽车的驾驶风格，以及其可识别性。受试者被随机分配至3个组别之一：防御型自动驾驶汽车组、激进型自动驾驶汽车组，以及混合型自动驾驶汽车组。受试者所属组别决定了其实验中遇到的自动驾驶汽车的驾驶风格。
更多关于实验设置的细节可参阅我们已发表的论文：https://doi.org/10.1016/j.trf.2022.09.018
**附带文件：** 处理后的数据集、说明文档（ReadMe）、数据处理脚本
**文件格式：** 数据文件采用.csv、.xlsx格式；Jupyter笔记本采用.ipynb格式