[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85918-en":3,"doc-seo-85918-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":13,"seo_description":14,"update_tm":28,"read_time":29},85918,7971461740886,"Theodore","https://ap-avatar.wpscdn.com/davatar_3d24733baf745e90a7e4bdd5f77d97b2",8,"Research & Report","Large Language Model Enhanced Differentiable Trajectory Planning for IoT-Enabled Autonomous Driving","Autonomous driving planning is central to IoT-enabled intelligent transportation systems, where vehicles must produce safe, efficient, and executable trajectories in complex urban settings using multi-source contextual information. Imitation learning can learn from large datasets, yet still struggles with poor long-tail interaction coverage, weak consistency with downstream constrained refinement, and limited high-level scene semantics under real-time constraints. A proposed LLM-enhanced differentiable planning framework adds surrounding-agent-centric augmentation, an asynchronous semantic enhancement module, and a differentiable residual-penalty optimizer for end-to-end refinement. Experiments achieve strong benchmark results and demonstrate real-time closed-loop deployment via CARLA-ROS tests.","Large Language Model Enhanced Differentiable Trajectory Planning  \nfor IoT-Enabled Autonomous Driving  \nShihao Zhang, Jing Yang, Ziyu Song, Zheng Lin, Sunil Prajapat, Zhaochen Xia, Hemant Ghayvat, Haitao Ding,  \nLip Yee Por, Ashok Kumar Das  \narXiv :2607 . 10438v1 [ cs .RO] 11 Jul 2026  \nAbstract—Autonomous driving planning is a key component of IoT-enabled intelligent transportation systems, requiring vehicles to generate safe, efficient, and executable trajectories in complex urban environments from multi-source contextual information. While imitation learning (IL) has shown promise on large-scale datasets, IL-based planners still suffer from limited coverage of complex long-tail interactions, weak consistency with downstream constrained refinement, and insufficient use of high level scene semantics under real time constraints. To address these issues, this paper proposes a large language model (LLM) enhanced differentiable trajectory planning framework for IoT-enabled autonomous driving. Specifically, we introduce a surrounding agent centric data augmentation strategy to reorganize surrounding agent trajectories as additional planning supervision, thereby improving the training distribution without collecting additional raw data. We further design a complexity-aware asynchronous LLM-based semantic enhancement module to extract scene-related high-level semantic features with controlled online overhead. In addition, a differentiable optimization module is incorporated to refine generated trajectories with explicit residual penalties while backpropagating optimization gradients to the upstream planner. Experiments show that the proposed method achieves the best overall scores of 83.63 and 78.29 on thenuPlan closed-loop nonreactive and reactive Hard20 benchmarks, respectively, and CARLA-ROS tests further verify its online deployment and real time closed-loop execution capability.  \nIndex Terms—Imitation Learning, Differentiable Optimization, Large Language Model, Trajectory Planning, Connected Autonomous Driving.  \nI. INTRODUCTION  \nThis work was supported by the China Automobile Industry Innovation and Development Joint Fund under Grant No. U1864206, and China Shenzhen Major Science and Technology Projects under Grant No. ZDCY20250901101003004 . (Corresponding author: Haitao Ding.)  \nShihao Zhang, Ziyu Song, Zhaochen Xia and Haitao Ding are with the State Key Laboratory of Automotive Simulation and Control, Jilin University, Changchun 130000, China (e-mail: [shihao24@mails.jlu.edu.cn](shihao24@mails.jlu.edu.cn); [songziyu@jlu.edu.cn](songziyu@jlu.edu.cn); [xiazc24@mails.jlu.edu.cn](xiazc24@mails.jlu.edu.cn); [dinght@jlu.edu.cn](dinght@jlu.edu.cn)).  \nJing Yang and Por Lip Yee are with the Center of Research for Cyber Security and Network (CSNET), Faculty of Computer Science and Information Technology, Universiti Malaya, 50603 Kuala Lumpur, Malaysia (e-mail: [s2147529@siswa.um.edu.my](s2147529@siswa.um.edu.my); [porlip@um.edu.my](porlip@um.edu.my)).  \nZheng Lin is with the Interdisciplinary Centre for Security, Reliability and Trust (SnT), University of Luxembourg, Luxembourg (e-mail: [zhenglin@ieee.org](zhenglin@ieee.org)).  \nSunil Prajapat and Hemant Ghayvat are with the IMT, Department of Humanities and Technology, Roskilde University, Roskilde, Denmark (e-mail: [sunilprajapat645@gmail.com](sunilprajapat645@gmail.com); [hghayvat@ruc.dk](hghayvat@ruc.dk)).  \nAshok Kumar Das is with the Center for Security, Theory and Algorithmic Research, International Institute of Information Technology, Hyderabad 500 032, India, and also with the Department of Computer Science and Engineering, College of Informatics, Korea University, 145 Anam-ro, Seongbuk-gu, Seoul 02841, South Korea (e-mail: [iitkgp.akdas@gmail.com](iitkgp.akdas@gmail.com), [ashok.das@iiit.ac.in](ashok.das@iiit.ac.in)) .  \nAUTONOMOUS driving systems are a key component  \nof IoT-enabled intelligent transportation services [1]–[4], where vehicles integrate multi-source cont","cbCairLE4vYBJpEq","https://ap.wps.com/l/cbCairLE4vYBJpEq","pdf",15130127,3,1,13,"English","en",105,"# Abstract\n# Introduction","[{\"question\":\"What problem does the paper address in imitation-learning based autonomous driving planning?\",\"answer\":\"It targets limited coverage of complex long-tail interactions, weak consistency with downstream constrained refinement, and insufficient high-level scene semantics under real-time constraints.\"},{\"question\":\"How does the proposed framework improve training without collecting additional raw data?\",\"answer\":\"It uses a surrounding agent centric data augmentation strategy that reorganizes surrounding agent trajectories as additional planning supervision, improving the training distribution.\"},{\"question\":\"What components are added to enhance and refine trajectories during planning?\",\"answer\":\"A complexity-aware asynchronous LLM-based semantic enhancement module extracts scene-related high-level semantics, and a differentiable optimization module refines trajectories using residual penalties with backpropagated optimization gradients to the upstream planner.\"}]",1784207164,33,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"large-language-model-enhanced-differentiable-trajectory-planning-for-iot-enabled-autonomous-driving","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,50],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":20},"https://docshare.wps.com/document/research-report/",{"item":51,"name":13,"@type":43,"position":52},"https://docshare.wps.com/document/large-language-model-enhanced-differentiable-trajectory-planning-for-iot-enabled-autonomous-driving/85918/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":24,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-24","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does the paper address in imitation-learning based autonomous driving planning?","Question",{"text":75,"@type":76},"It targets limited coverage of complex long-tail interactions, weak consistency with downstream constrained refinement, and insufficient high-level scene semantics under real-time constraints.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does the proposed framework improve training without collecting additional raw data?",{"text":80,"@type":76},"It uses a surrounding agent centric data augmentation strategy that reorganizes surrounding agent trajectories as additional planning supervision, improving the training distribution.",{"name":82,"@type":73,"acceptedAnswer":83},"What components are added to enhance and refine trajectories during planning?",{"text":84,"@type":76},"A complexity-aware asynchronous LLM-based semantic enhancement module extracts scene-related high-level semantics, and a differentiable optimization module refines trajectories using residual penalties with backpropagated optimization gradients to the upstream planner.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]