How should autonomous vehicles drive? Policy, methodological, and social considerations for designing a driver

Author:

Bin-Nun Amitai Y.ORCID,Derler Patricia,Mehdipour Noushin,Tebbens Radboud Duintjer

Abstract

AbstractAutonomous vehicles (AVs) are being developed, tested, and commercialized around the world. While skilled human drivers can rely on their experience and common sense to navigate complex driving situations that involve trade-offs between competing objectives, AVs are engineered systems, which may handle complex scenarios based on driving principles articulated at the time of system design. This raises the question of what constitutes proper driving behavior in a complex driving scenario. Many jurisdictions point to existing rules of the road as a description of good driving and, by requiring AVs to follow such rules, hope to improve the safety and efficiency of the transportation system. This paper discusses the desirability of a comprehensive definition of AV behavior, reviews subnational, national, and international regulatory developments that seek to define how AVs might drive, and discusses the tensions between safe, lawful, and efficient driving. Locally defined rules of the road can serve as a guide to a comprehensive driving behavior specification. However, translating rules of the road, which are legal documents written in natural language, to formal rules for use by computers deployed on AVs is a challenging task. In particular, the pervasive appeals to judgment that are present in many rules of the road do not easily lend themselves to the precise formalization of conditions and quantification of values that computers use to make decisions. This work also considers the effect that formalizing behavior for adoption by AVs might have on the general driving culture, and especially on the relationship between existing classes of road users. To highlight the challenges associated with formalizing the rules of the road, this work reports on an experiment where two teams independently translated two rules of the road into formal rules to instruct AVs or to verify the correctness of AV behavior. The study results emphasize the desirability of new technical and political structures to mediate a shared understanding of the rules of the road. The harmonization of behavioral expectations has the potential to improve the safety and efficiency of AV deployments, as well as the broader transportation system.

Publisher

Springer Science and Business Media LLC

Subject

General Economics, Econometrics and Finance,General Psychology,General Social Sciences,General Arts and Humanities,General Business, Management and Accounting

Reference94 articles.

1. Allen BL, Shin BT, Cooper PJ (1978) Analysis of traffic conflicts and collisions. Transp Res Board (667):67–74

2. Alliance for Automotive Innovation (2022) Alliance for automotive innovation, policy roadmap to advance automated vehicle innovation. https://www.autosinnovate.org/innovation/AVRoadmap.pdf. Accessed Jun 2022

3. Anesiadou A, Makridis M, Mattas K, Fontaras G, Ciuffo B (2021) Characterization of drivers heterogeneity and its integration within traffic simulation. https://arxiv.org/abs/2107.02618

4. Archer J, Young W (2010) The measurement and modelling of proximal safety measures. In: Proceedings of the Institution of Civil Engineers-Transport vol 163, pp. 191–201. Thomas Telford Ltd

5. Arechiga N (2019) Specifying safety of autonomous vehicles in signal temporal logic. In: 2019 IEEE Intelligent Vehicles Symposium (IV). IEEE, 2019. pp. 58–63

Cited by 2 articles. 订阅此论文施引文献 订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3