Yang Guan, Assistant Researcher, School of Vehicle and Mobility, Tsinghua University. He has long been committed to the research of brain-inspired learning systems and end-to-end models for high-level autonomous driving vehicles, and has put forward the following achievements: (1) An integrated decision-making and control architecture for autonomous vehicles, which achieves the dual goals of efficient model iteration and real-time online computing. This architecture supported the open-road testing of China's first full-stack end-to-end autonomous driving system, was reported on the official website of Tsinghua University, and has been deployed and applied in enterprises such as GAC, Toyota, Didi, and Meituan; (2) A direct reinforcement learning algorithm driven by a hybrid of data and models, which overcomes the constraints of poor training performance and slow convergence speed. This algorithm was awarded the First Global Outstanding Project Award for Industry-University-Research Cooperation of Didi in 2022, and has been applied to the training of Tsinghua iDrive end-to-end models; (3) A reinforcement fine-tuning method for end-to-end models combined with real vehicle data, which solves the problem of low reliability of supervised pre-trained models. This method supported the stable operation of Meituan's automatic delivery vehicles for more than 5 million kilometers, and contributed to the first pilot achievement of the Transportation Power China initiative in the domestic automatic delivery field. He has published 19 papers in total and participated in compiling 1 monograph. Among them, 7 papers were published as the first author (including co-first author), including 1 cover paper, 6 papers with an impact factor greater than 10, and 5 papers won Excellent Paper Awards.
【Grants】
(1) National Natural Science Foundation of China, Major Research Plan, No.20251310051, Construction method and iterative evolution of vehicle–road–cloud
integrated end-to-end autonomous driving functional software, Jan 2026 - Dec 2029, Funding: RMB 3,851,400, Ongoing, Participant
(2) Dongfeng Motor Group Co., Ltd., Horizontal Project, No.20252002249, Development Project of Autonomous Driving Large Model Algorithm Based on VLA, Dec 2025 - Oct 2026, Funding: RMB 2,997,300, Ongoing, Principal Investigator
(3) Beijing Sankuai Online Technology Co., Ltd., Horizontal Project, No.20252930041, Training of Multi-modal End-to-end Autonomous Driving Model and Reinforcement Learning Fine-tuning, Nov 2025 - Dec 2027, Funding: RMB 686,000, Ongoing, Principal Investigator
