Autonomous vehicle safety: Understanding perceptions of pedestrians and bicyclists |
| |
| Affiliation: | 1. Charité – Universitätsmedizin Berlin, Corporate Member of Freie Universität Berlin and Humboldt-Universität zu Berlin, Institute of Medical Sociology and Rehabilitation Science, Charitéplatz 1, 10117 Berlin, Germany;2. Karlsruher Institute of Technology, Institut für Arbeitswissenschaft und Betriebsorganisation, Engler-Bunte-Ring 4, 76131 Karlsruhe, Germany;3. Freie Universität Berlin, Institute of Philosophy, Habelschwerdter Allee 30, 14195 Berlin, Germany;4. Humboldt-Universität zu Berlin, Institute of Social Sciences, Universitätsstr. 3b, 10117 Berlin, Germany;1. Department of Industrial & Systems Engineering, Mississippi State University, PO Box 9542, MS 39762, USA;2. Center for Advanced Vehicular Systems, Mississippi State University, PO Box 5405, MS 39762, USA;1. AKKA Technologies, Guyancourt, France;2. Research Department, Renault SAS, Guyancourt, France;1. The University of Texas at Austin, Department of Civil, Architectural and Environmental Engineering, 301 E. Dean Keeton St. Stop C1761, Austin, TX 78712, USA;2. The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong |
| |
| Abstract: | Autonomous vehicle (AV) technologies have been rapidly advancing. One benefit of AVs is that the technology could eliminate many driver errors and also mitigate many pedestrian and bicyclist collisions. Real-world AVs have been tested in many cities. Five companies are running around 50 AVs in Pittsburgh, following the autonomous testing guidelines. BikePGH, a non-profit organization located in Pittsburgh, Pennsylvania conducted a follow-up survey in 2019 (the first survey was conducted in 2017) to understand non-motorists’ opinions of AVs. This study examined how pedestrians and bicyclists perceived AV safety based on their understanding and experiences. At first, this study performed a comparison group test to determine which questions vary by participants’ AV safety rating. The responses were later analyzed with a data mining method known as ‘association rules mining.’ A new performance measure, known as the rule power factor, was then used to identify the significant patterns in the form of rules. The participants also provided their thoughts in responses to the open-ended questions. Using Latent Dirichlet Allocation (LDA), a topic modeling algorithm, 40 topic models were developed based on five open-ended questions. The findings show that the non-motorists showed comparatively fewer negative opinions towards AVs than positive assessments. The results also show that perception patterns vary by the participant’s rating on AV safety. Findings of this study would be beneficial for the AV stakeholders in making AVs and roadways safer for non-motorists. |
| |
| Keywords: | Autonomous vehicles Pedestrians Bicyclists Perception Association rules mining Topic modeling |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
