Jingda Wu , Chao Huang , Hailong Huang , Chen Lv , Yuntong Wang , Fei-Yue Wang
{"title":"基于强化学习的自动驾驶行为规划的最新进展:调查","authors":"Jingda Wu , Chao Huang , Hailong Huang , Chen Lv , Yuntong Wang , Fei-Yue Wang","doi":"10.1016/j.trc.2024.104654","DOIUrl":null,"url":null,"abstract":"<div><p>Autonomous driving (AD) holds the potential to revolutionize transportation efficiency, but its success hinges on robust behavior planning (BP) mechanisms. Reinforcement learning (RL) emerges as a pivotal tool in crafting these BP strategies. This paper offers a comprehensive review of RL-based BP strategies, spotlighting advancements from 2021 to 2023. We completely organize and distill the relevant literature, emphasizing paradigm shifts in RL-based BP. Introducing a novel categorization, we trace the trajectory of efforts aimed at surmounting practical challenges encountered by autonomous vehicles through innovative RL techniques. To guide readers, we furnish a quantitative analysis that maps the volume and diversity of recent RL configurations, elucidating prevailing trends. Additionally, we delve into the imminent challenges and potential directions for the future of RL-driven BP in AD. These directions encompass addressing safety vulnerabilities, fostering continual learning capabilities, enhancing data efficiency, championing collaborative vehicular cloud networks, integrating large language models, and enhancing ethical considerations.</p></div>","PeriodicalId":54417,"journal":{"name":"Transportation Research Part C-Emerging Technologies","volume":null,"pages":null},"PeriodicalIF":7.6000,"publicationDate":"2024-05-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Recent advances in reinforcement learning-based autonomous driving behavior planning: A survey\",\"authors\":\"Jingda Wu , Chao Huang , Hailong Huang , Chen Lv , Yuntong Wang , Fei-Yue Wang\",\"doi\":\"10.1016/j.trc.2024.104654\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><p>Autonomous driving (AD) holds the potential to revolutionize transportation efficiency, but its success hinges on robust behavior planning (BP) mechanisms. Reinforcement learning (RL) emerges as a pivotal tool in crafting these BP strategies. This paper offers a comprehensive review of RL-based BP strategies, spotlighting advancements from 2021 to 2023. We completely organize and distill the relevant literature, emphasizing paradigm shifts in RL-based BP. Introducing a novel categorization, we trace the trajectory of efforts aimed at surmounting practical challenges encountered by autonomous vehicles through innovative RL techniques. To guide readers, we furnish a quantitative analysis that maps the volume and diversity of recent RL configurations, elucidating prevailing trends. Additionally, we delve into the imminent challenges and potential directions for the future of RL-driven BP in AD. These directions encompass addressing safety vulnerabilities, fostering continual learning capabilities, enhancing data efficiency, championing collaborative vehicular cloud networks, integrating large language models, and enhancing ethical considerations.</p></div>\",\"PeriodicalId\":54417,\"journal\":{\"name\":\"Transportation Research Part C-Emerging Technologies\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":7.6000,\"publicationDate\":\"2024-05-22\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Transportation Research Part C-Emerging Technologies\",\"FirstCategoryId\":\"5\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S0968090X2400175X\",\"RegionNum\":1,\"RegionCategory\":\"工程技术\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"TRANSPORTATION SCIENCE & TECHNOLOGY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Transportation Research Part C-Emerging Technologies","FirstCategoryId":"5","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0968090X2400175X","RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"TRANSPORTATION SCIENCE & TECHNOLOGY","Score":null,"Total":0}
引用次数: 0
摘要
自动驾驶(AD)有望彻底改变交通效率,但其成功与否取决于强大的行为规划(BP)机制。强化学习(RL)是制定这些 BP 策略的关键工具。本文全面回顾了基于 RL 的 BP 策略,重点介绍了从 2021 年到 2023 年的进展。我们对相关文献进行了全面整理和提炼,强调了基于 RL 的 BP 的范式转变。我们引入了一种新颖的分类方法,追溯了旨在通过创新 RL 技术克服自动驾驶汽车遇到的实际挑战的努力轨迹。为了给读者提供指导,我们提供了定量分析,描绘了近期 RL 配置的数量和多样性,阐明了当前的趋势。此外,我们还深入探讨了RL驱动的自动驾驶汽车BP未来即将面临的挑战和潜在发展方向。这些方向包括解决安全漏洞、培养持续学习能力、提高数据效率、支持协作式车载云网络、整合大型语言模型以及加强道德考量。
Recent advances in reinforcement learning-based autonomous driving behavior planning: A survey
Autonomous driving (AD) holds the potential to revolutionize transportation efficiency, but its success hinges on robust behavior planning (BP) mechanisms. Reinforcement learning (RL) emerges as a pivotal tool in crafting these BP strategies. This paper offers a comprehensive review of RL-based BP strategies, spotlighting advancements from 2021 to 2023. We completely organize and distill the relevant literature, emphasizing paradigm shifts in RL-based BP. Introducing a novel categorization, we trace the trajectory of efforts aimed at surmounting practical challenges encountered by autonomous vehicles through innovative RL techniques. To guide readers, we furnish a quantitative analysis that maps the volume and diversity of recent RL configurations, elucidating prevailing trends. Additionally, we delve into the imminent challenges and potential directions for the future of RL-driven BP in AD. These directions encompass addressing safety vulnerabilities, fostering continual learning capabilities, enhancing data efficiency, championing collaborative vehicular cloud networks, integrating large language models, and enhancing ethical considerations.
期刊介绍:
Transportation Research: Part C (TR_C) is dedicated to showcasing high-quality, scholarly research that delves into the development, applications, and implications of transportation systems and emerging technologies. Our focus lies not solely on individual technologies, but rather on their broader implications for the planning, design, operation, control, maintenance, and rehabilitation of transportation systems, services, and components. In essence, the intellectual core of the journal revolves around the transportation aspect rather than the technology itself. We actively encourage the integration of quantitative methods from diverse fields such as operations research, control systems, complex networks, computer science, and artificial intelligence. Join us in exploring the intersection of transportation systems and emerging technologies to drive innovation and progress in the field.