Jiahui Xu;Wenbo Shao;Weida Wang;Cheng Liu;Chao Yang;Jun Li;Hong Wang
{"title":"From Component to System: A Task-Unified Planning System With Planning-Oriented Predictor","authors":"Jiahui Xu;Wenbo Shao;Weida Wang;Cheng Liu;Chao Yang;Jun Li;Hong Wang","doi":"10.1109/TVT.2024.3519178","DOIUrl":null,"url":null,"abstract":"Autonomous driving is developing rapidly and has become a hot topic in both industry and research. The planning system plays a crucial role in meeting the requirements of autonomous driving. However, current planning system designs may not effectively serve planning tasks. A typical modular planning system offers high interpretability and flexibility. However, it may cause task-agnostic problems between the upstream predictor and the downstream planner. End-to-end driving systems have a natural advantage in achieving system-wide integration, but their poor interpretability poses safety risks. Therefore, in this paper, a task-unified planning framework is proposed to inspire the current prediction-planning paradigm. In this architecture, driving tasks are first modeled. Then, the predictor and planner are jointly designed and optimized based on these tasks. Finally, during the actual planning process, the upstream and downstream components remain relatively independent to allow for flexible adjustments. The core of this architecture is a planning-oriented predictor named POP, which fully retains the advantages of modular systems by optimizing the predictor to meet driving requirements. Comprehensive experiments demonstrate its effectiveness. Compared to typical modular systems, POP-based framework shows significant improvements in planning tasks, particularly in collision avoidance, ensuring system safety without compromising driving efficiency or comfort.","PeriodicalId":13421,"journal":{"name":"IEEE Transactions on Vehicular Technology","volume":"74 4","pages":"5335-5348"},"PeriodicalIF":7.1000,"publicationDate":"2024-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"IEEE Transactions on Vehicular Technology","FirstCategoryId":"94","ListUrlMain":"https://ieeexplore.ieee.org/document/10804683/","RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENGINEERING, ELECTRICAL & ELECTRONIC","Score":null,"Total":0}
引用次数: 0
Abstract
Autonomous driving is developing rapidly and has become a hot topic in both industry and research. The planning system plays a crucial role in meeting the requirements of autonomous driving. However, current planning system designs may not effectively serve planning tasks. A typical modular planning system offers high interpretability and flexibility. However, it may cause task-agnostic problems between the upstream predictor and the downstream planner. End-to-end driving systems have a natural advantage in achieving system-wide integration, but their poor interpretability poses safety risks. Therefore, in this paper, a task-unified planning framework is proposed to inspire the current prediction-planning paradigm. In this architecture, driving tasks are first modeled. Then, the predictor and planner are jointly designed and optimized based on these tasks. Finally, during the actual planning process, the upstream and downstream components remain relatively independent to allow for flexible adjustments. The core of this architecture is a planning-oriented predictor named POP, which fully retains the advantages of modular systems by optimizing the predictor to meet driving requirements. Comprehensive experiments demonstrate its effectiveness. Compared to typical modular systems, POP-based framework shows significant improvements in planning tasks, particularly in collision avoidance, ensuring system safety without compromising driving efficiency or comfort.
期刊介绍：
The scope of the Transactions is threefold (which was approved by the IEEE Periodicals Committee in 1967) and is published on the journal website as follows: Communications: The use of mobile radio on land, sea, and air, including cellular radio, two-way radio, and one-way radio, with applications to dispatch and control vehicles, mobile radiotelephone, radio paging, and status monitoring and reporting. Related areas include spectrum usage, component radio equipment such as cavities and antennas, compute control for radio systems, digital modulation and transmission techniques, mobile radio circuit design, radio propagation for vehicular communications, effects of ignition noise and radio frequency interference, and consideration of the vehicle as part of the radio operating environment. Transportation Systems: The use of electronic technology for the control of ground transportation systems including, but not limited to, traffic aid systems; traffic control systems; automatic vehicle identification, location, and monitoring systems; automated transport systems, with single and multiple vehicle control; and moving walkways or people-movers. Vehicular Electronics: The use of electronic or electrical components and systems for control, propulsion, or auxiliary functions, including but not limited to, electronic controls for engineer, drive train, convenience, safety, and other vehicle systems; sensors, actuators, and microprocessors for onboard use; electronic fuel control systems; vehicle electrical components and systems collision avoidance systems; electromagnetic compatibility in the vehicle environment; and electric vehicles and controls.
京公网安备 11010802042870号
京ICP备2023020795号-1
分享
求助内容：
应助结果提醒方式：
扫码关注我们
