Investigating collision patterns to support autonomous driving safety

Lee, Carmen Kar Hang, Leung, Eric Ka Ho, Tse, Ying ORCID: https://orcid.org/0000-0001-6174-0326 and Tsao, Yu-Chung 2024. Investigating collision patterns to support autonomous driving safety. Enterprise Information Systems 18 (2) , 2243460. 10.1080/17517575.2023.2243460 Item availability restricted.

PDF - Accepted Post-Print Version Restricted to Repository staff only until 2 August 2024 due to copyright restrictions. Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (531kB)

Abstract

There is a debate on the importance of autonomous vehicles (AVs) and the methods for ensuring that AVs are safe to be driven on roads open to traffic, which has become the priority for engineers, artificial intelligence experts, and data scientists who wish to enhance autonomous driving safety. This paper proposes a machine learning framework for investigating AV collisions, thereby advancing the knowledge on the risk factors of AV collisions. We used the California Department of Motor Vehicles’ AV collision reports from January, 2019 to October, 2021 to determine the association between risk factors and the level of damage to an AV due to collisions. Association rule mining was used to develop methodologies that can advance result interpretability, which is crucial in the transportation field as it will lead to the development of evidence-based policies. A total of twenty-one rules were determined and used to reveal the unique safety patterns of AVs to understand the factors that co-occur with AV damage. This study demonstrates that collision data, when analysed using appropriate machine learning algorithms, can generate useful insights that complement current AV-related policies and provide practical information that can be used to enhance autonomous driving safety in the long term.

Item Type:	Article
Date Type:	Publication
Status:	Published
Schools:	Business (Including Economics)
Publisher:	Taylor and Francis Group
ISSN:	1751-7575
Date of First Compliant Deposit:	6 September 2023
Date of Acceptance:	29 July 2023
Last Modified:	20 Mar 2024 16:37
URI:	https://orca.cardiff.ac.uk/id/eprint/162276

Actions (repository staff only)

Edit Item