Cheng Chang

I am a 5th year Ph.D. Candidate in the Department of Automation, Tsinghua University, advised by Prof. Li Li . In 2021, I obtained my B.S. degree in the Department of Automation, Tsinghua University.

I am broadly interested in LLM and autonomous driving. My current research focuses on LLM driven scenario understanding, and multi-agent applications (prediction, planning, warning, simulation, etc.) for intelligent systems.

Email / Google Scholar / ResearchGate / Github

News

2025-05: The journal paper Driving-RAG is accepted by Automotive Innovation.

2024-05: The journal paper VistaScenario is published on TIV (Full-Length Paper).

2024-04: The journal paper LLMScenario is published on TSMC, also appears on Popular Articles Column.

2023-12: The journal version of CAV Warning is published on ENGINEERING Management (Journal of Chinese Academy of Engineering).

2023-07: The journal paper BEV-V2X is published on TIV (Full-Length Paper), also appears on Popular Articles Column.

2022-10: The journal paper MetaScenario is published on TIV (Full-Length Paper).

2022-10: The conference paper CAV Warning is honored as the Best Student Paper Award at ITSC 2022!

2022-06: The conference paper CAV Warning is accepted to ITSC 2022.

First Author

	MetaScenario: A Framework for Driving Scenario Data Description, Storage and Indexing Cheng Chang, Dongpu Cao, Long Chen, Kui Su, Kuifeng Su, Yuelong Su, Fei-Yue Wang, Jue Wang, Ping Wang, Junqing Wei, Gansha Wu, Xiangbin Wu, Huile Xu, Nanning Zheng, Li Li IEEE Transactions on Intelligent Vehicles (TIV), 2023 [IEEE] [ResearchGate] [Code] In cooperation with many famous universities (XJTU, UCAS, PKU) and companies (Alibaba, Tencent, AutoNavi, DiDi, UISEE, Intel, Momenta), we propose MetaScenario, a unified and efficient framework for driving scenario data processing, storage, and indexing.
	BEV-V2X: Cooperative Birds-Eye-View Fusion and Grid Occupancy Prediction via V2X-Based Data Sharing Cheng Chang, Jiawei Zhang, Kunpeng Zhang, Wenqin Zhong, Xinyu Peng, Shen Li, Li Li IEEE Transactions on Intelligent Vehicles (TIV), 2023 [IEEE] [ResearchGate] BEV perception data constructed by single vehicles encounter certain issues, such as low accuracy and insufficient range. BEV-V2X utilizes V2X technique and Attention-based network to comprehensively fuse and predict the driving scenarios.
	LLMScenario: Large Language Model Driven Scenario Generation Cheng Chang, Siqi Wang, Jiawei Zhang, Jingwei Ge, Li Li IEEE Transactions on Systems, Man, and Cybernetics: Systems (TSMC), 2024 [IEEE] [ResearchGate] The LLMScenario framework, which consists of scenario prompt engineering, LLM core, and evaluation feedback tuning, provides the potential to understand and generate safety-critical scenarios via LLM.
	Driving-RAG: Driving Scenarios Embedding, Search, and RAG Applications Cheng Chang, Jingwei Ge, Jiazhe Guo, Zelin Guo, Binghong Jiang, Li Li Automotive Innovation (AUIN), 2025 [Arxiv] [自动驾驶之心] Driving-RAG addresses the challenges of efficient scenario data embedding, search, and applications for RAG systems. Especially HNSW-TSD algorithm performs efficient vector search to achieve high efficiency without sacrificing accuracy. The reorganization by graph knowledge enhances the relevance to the prompt and augment LLM generation.
	Driving Safety Monitoring and Warning for Connected and Automated Vehicles via Edge Computing Cheng Chang, Kunpeng Zhang, Jiawei Zhang, Shen Li, Li Li IEEE International Conference on Intelligent Transportation Systems (ITSC), 2022 [IEEE] [ResearchGate] With edge computing system, we propose different fast algorithms and corresponding data structure models to calculate risks based on timely received data for different types of CAVs.
	CAV Driving Safety Monitoring and Warning via V2X-Based Edge Computing System Cheng Chang, Jiawei Zhang, Kunpeng Zhang, Yichen Zheng, Mengkai Shi, Jianming Hu, Shen Li, Li Li ENGINEERING Management (EM), 2024 [Springer] [ResearchGate] In the journal version, we examine the prerequisites for data accuracy and transmission delay to enhance the safety of CAV driving. The framework is validated through both simulated and real-world road tests, proving its utility in diverse driving conditions.
	VistaScenario: Interaction Scenario Engineering for Vehicles with Intelligent Systems for Transport Automation Cheng Chang, Jiawei Zhang, Jingwei Ge, Zuo Zhang, Junqing Wei, Li Li, Fei-Yue Wang IEEE Transactions on Intelligent Vehicles (TIV), 2024 [IEEE] [ResearchGate] A novel scenario metric Graph-DTW is proposed to conduct scenario comparison and labeling, which efficiently extracts the extreme corner cases in scenario database.

Co-Author

* indicates equal contribution

	Mixing Left and Right-Hand Driving Data in a Hierarchical Framework With LLM Generation Jiazhe Guo, Cheng Chang, Zhiheng Li, Li Li IEEE Robotics and Automation Letters (RAL), 2024 [IEEE] [ResearchGate]
	A Systematic Solution of Human Driving Behavior Modeling and Simulation for Automated Vehicle Studies Kunpeng Zhang, Cheng Chang, Wenqin Zhong, Shen Li, Zhiheng Li, Li Li IEEE Transactions on Intelligent Transportation Systems (TITS), 2022 [IEEE] [ResearchGate]
	LLM-based Operating Systems for Automated Vehicles: A New Perspective Jingwei Ge, Cheng Chang, Jiawei Zhang, Lingxi Li, Xiaoxiang Na, Yilun Lin, Li Li, Fei-Yue Wang IEEE Transactions on Intelligent Vehicles (TIV) (Perspectives), 2024 [IEEE] [ResearchGate] This paper envisions a revolution of the LLM based Intelligent Operating Systems to support the core of automated vehicles.
	CAVSim: A Microscopic Traffic Simulator for Evaluation of Connected and Automated Vehicles Jiawei Zhang, Cheng Chang, Zimin He, Wenqin Zhong, Danya Yao, Shen Li, Li Li IEEE Transactions on Intelligent Transportation Systems (TITS), 2023 [IEEE] [ResearchGate] [Code]
	Multi-Agent DRL-Based Lane Change With Right-of-Way Collaboration Awareness Jiawei Zhang, Cheng Chang, Xianlin Zeng, Li Li IEEE Transactions on Intelligent Transportation Systems (TITS), 2022 [IEEE] [ResearchGate]
	A Generic Optimization-Based Enhancement Method for Trajectory Data: Two Plus One Feng Zhu, Cheng Chang, Zhiheng Li, Boqi Li, Li Li Accident Analysis & Prevention (AAP), 2024 [Elsevier] [ResearchGate]
	CD-DB: A Data Storage Model for Cooperative Driving Haiyang Yu, Cheng Chang, Shen Li, Li Li IEEE Transactions on Intelligent Vehicles (TIV), 2022 [IEEE] [ResearchGate]
	Robust Multitask Learning With Sample Gradient Similarity Xinyu Peng, Cheng Chang, Fei-Yue Wang, Li Li IEEE Transactions on Systems, Man, and Cybernetics: Systems (TSMC), 2024 [IEEE] [ResearchGate]
	Task-Driven Controllable Scenario Generation Framework Based on AOG Jingwei Ge, Jiawei Zhang, Cheng Chang, Yi Zhang, Danya Yao, Li Li IEEE Transactions on Intelligent Transportation Systems (TITS), 2023 [IEEE] [ResearchGate]
	Dynamic Testing for Autonomous Vehicles Using Random Quasi Monte Carlo Jingwei Ge, Jiawei Zhang, Cheng Chang, Yi Zhang, Danya Yao, Yonglin Tian, Li Li IEEE Transactions on Intelligent Vehicles (TIV), 2023 [IEEE] [ResearchGate]
	Life-long Learning and Testing for Automated Vehicles via Adaptive Scenario Sampling as A Continuous Optimization Process Jingwei Ge, Pengbo Wang, Cheng Chang, Yi Zhang, Danya Yao, Li Li IEEE Transactions on Intelligent Vehicles (TIV), 2024 [IEEE] [ResearchGate]
	Unleashing the Two-Dimensional Benefits of Connected and Automated Vehicles via Dedicated Intersections in Mixed Traffic Jiawei Zhang, Cheng Chang, Shen Li, Xuegang(Jeff) Ban, Li Li Transportation Research Part C:Emerging Technologies (TR-C), 2024 [Elsevier] [ResearchGate]