A review of best practices, standards, and approaches for transportation safety data and driver state prediction

Author:

Nartey David1ORCID,Alambeigi Hananeh1,McDonald Anthony D.2ORCID,Shipp Eva3,Manser Michael3,Christensen Scott4,Lenneman John K.5,Pulver Elizabeth4

Affiliation:

1. Texas A&M University, College Station, TX, USA

2. University of Wisconsin-Madison, Madison, WI, USA

3. Texas A&M Transportation Institute, Bryan, TX, USA

4. State Farm Mutual Auto Insurance Co., Bloomington, IL, USA

5. Toyota Collaborative Safety Research Center, Ann Arbor, MI

Abstract

This systematic review documents current best practices, standards, and approaches for transportation safety data analytics. While standards exist for defining measures, there are few available standards or guides for processing driving and driver data. Standards are crucial for ensuring repeatability and appropriate cost-benefit decisions. The review identified 36 relevant studies describing behavioral and physiological measures. Most studies do not comprehensively report data processing steps. Of the studies that did report data processing steps, few analyzed the impact of decisions made during data processing on algorithm performance. Most studies were conducted in a controlled simulator environment and may not generalize to naturalistic settings. The findings show that driver behavior and physiological data show efficacy for detecting fatigue, distraction, stress, and driver errors. The results of these studies may necessitate additional data processing standards and future work should focus on measuring the effects of data decisions on model performance.

Funder

Toyota Collaborative Safety Research Center

State Farm

National Highway Traffic Safety Administration

Publisher

SAGE Publications

Subject

General Medicine,General Chemistry

Reference43 articles.

1. Alambeigi H., McDonald A. D., Manser M., et al. (2021). Predicting Driver Errors During Automated Vehicle Takeovers. Transportation Research Record (In Press).

2. Development and interval testing of a naturalistic driving methodology to evaluate driving behavior in clinical research

3. Physiological correlates of discomfort in automated driving

4. Driver workload and eye blink duration

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