Standard Grant: Ethical Algorithms in Autonomous Vehicles

标准拨款：自动驾驶汽车中的道德算法

• 批准号: 1734521
• 负责人: Nicholas Evans
• 金额: $ 55.67万
• 依托单位: University of Massachusetts Lowell
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2017
• 资助国家: 美国
• 起止时间: 2017-07-01 至 2022-06-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1734521&HistoricalAwards=false
• 关键词: Standard; Grant; Ethical; Algorithms; Autonomous

This project supports a team of researchers, including experts in moral philosophy and transportation engineering, who plan to address ethical issues that arise in the development of computer algorithms for autonomous (self-driving) vehicles. The project will address concerns over the expression of ethical values in self-driving vehicle when (for example) the vehicle detects an imminent and unavoidable crash and must select among navigation options, such as colliding with a crowded bus or with a lone motorcyclist. It aims to develop algorithmic representations of ethical decision-making frameworks for autonomous vehicles, and then model these frameworks over a range of contexts for their effect on risk management in traffic networks and for their expected public health impact. The overarching aim of the project is to promote principled approaches to ethical decision-making in complex, autonomous systems, and to prepare designers and engineers to encode ethical behavior into such systems. Research mentoring is to be coordinated through the host institution that will enable undergraduate and graduate engineers to participate in the project. A summer intensive will bring together students from diverse backgrounds to work with the project team and invited scholars. A postdoctoral fellow will conduct research on the ethics of autonomous vehicles, and assist in project administration.The project has two specific aims. It aims to develop ethical algorithms that can be converted into computer code for use in autonomous vehicles. Team members will review the literature on existing decision-making for autonomous vehicles, and then develop conceptual, algorithmic representations of ethical behavior that can be converted into computer code. They will also model changes in traffic risk profiles and expected public health impact of these algorithms; specifically, they will apply injury severity and crash frequency modeling using algorithms developed in meeting the first aim in a range of traffic contexts and market penetration scenarios. They will also develop a model for the expected public health benefits from these algorithms using model of resource and time-based triaging as a starting point. Both aims support training of scholars and practitioners sensitive to the ethical implications of autonomous vehicles. These pedagogical aspects have been designed to promote diverse interactions between STEM students and practitioners, and they will serve to improve STEM education and educator development.

该项目支持一个研究团队，包括道德哲学和交通工程专家，他们计划解决自动驾驶汽车计算机算法开发中出现的道德问题。该项目将解决人们对自动驾驶汽车道德价值观表达的担忧，例如，当车辆检测到即将发生且不可避免的碰撞时，必须在导航选项中进行选择，例如与拥挤的公共汽车或孤独的摩托车手相撞。它旨在开发自动驾驶汽车道德决策框架的算法表示，然后在一系列背景下对这些框架进行建模，以了解它们对交通网络风险管理的影响以及预期的公共健康影响。该项目的总体目标是在复杂的自主系统中推广道德决策的原则性方法，并帮助设计师和工程师将道德行为编码到此类系统中。研究指导将通过主办机构进行协调，使本科生和研究生工程师能够参与该项目。夏季强化课程将汇集来自不同背景的学生，与项目团队和受邀学者一起工作。一名博士后研究员将开展自动驾驶汽车伦理研究，并协助项目管理。该项目有两个具体目标。它的目标是开发可转换为自动驾驶汽车使用的计算机代码的道德算法。团队成员将回顾有关自动驾驶汽车现有决策的文献，然后开发可以转换为计算机代码的道德行为的概念算法表示。他们还将对交通风险状况的变化以及这些算法的预期公共健康影响进行建模；具体来说，他们将使用为满足一系列交通环境和市场渗透场景中的第一个目标而开发的算法来应用伤害严重程度和碰撞频率建模。他们还将使用资源和基于时间的分类模型作为起点，开发一个模型，以从这些算法中获得预期的公共健康效益。这两个目标都支持对对自动驾驶汽车的道德影响敏感的学者和从业者进行培训。这些教学方面的目的是促进 STEM 学生和从业者之间的多样化互动，并将有助于改善 STEM 教育和教育者发展。

项目成果

Reinforcement Learning based cooperative longitudinal control for reducing traffic oscillations and improving platoon stability

• DOI: 10.1016/j.trc.2022.103744
• 发表时间: 2022-06-09
• 期刊: TRANSPORTATION RESEARCH PART C-EMERGING TECHNOLOGIES
• 影响因子: 8.3
• 作者: Jiang, Liming;Xie, Yuanchang;Chen, Danjue
• 通讯作者: Chen, Danjue

Cooperative Car-Following and Merging: A Novel Merge Control Strategy Considering Cooperative Adaptive Cruise Control and Courtesy

协作跟车与并道：一种考虑协作自适应巡航控制和礼让的新型并道控制策略

• 发表时间: 2021
• 期刊: Proceedings of the 100th Transportation Research Board Annual Meeting
• 作者: Jiang, L.

Cooperative Highway Work Zone Merge Control Based on Reinforcement Learning in a Connected and Automated Environment

• DOI: 10.1177/0361198120935873
• 发表时间: 2020-07-17
• 期刊: TRANSPORTATION RESEARCH RECORD
• 影响因子: 1.7
• 作者: Ren, Tianzhu;Xie, Yuanchang;Jiang, Liming
• 通讯作者: Jiang, Liming

Dampen the Stop-and-Go Traffic with Connected and Automated Vehicles – A Deep Reinforcement Learning Approach*

• DOI: 10.1109/mt-its49943.2021.9529289
• 发表时间: 2020-05
• 期刊: 2021 7th International Conference on Models and Technologies for Intelligent Transportation Systems (MT-ITS)
• 作者: Liming Jiang;Yuanchang Xie;Danjue Chen;Tienan Li;Nicholas G. Evans

New England merge: a novel cooperative merge control method for improving highway work zone mobility and safety

• DOI: 10.1080/15472450.2020.1822747
• 发表时间: 2020-09-22
• 期刊: JOURNAL OF INTELLIGENT TRANSPORTATION SYSTEMS
• 影响因子: 3.6
• 作者: Ren, Tianzhu;Xie, Yuanchang;Jiang, Liming
• 通讯作者: Jiang, Liming