Hi! I am a postgraduate student with the School of Transportation Science and Engineering, Key Laboratory of Autonomous Transportation Technology for Special Vehicles, Ministry of Industry and Information Technology, Beihang University, working under the supervision of Professor Peng Chen. Before that, I got my Bachelor degree at Beihang University.

My research interests cover multi-agent interaction modeling in traffic flow and integration of prediction and planning for intelligent vehicles. I am actively looking for PhD positions in autonomous vehicle and embodied AI, with a particular interest in generative network world model, multimodal learning.

📝 Publications

Highlights

RA-L

Learning Rollout from Sampling: An R1-Style Tokenized Traffic Simulation Model

Ziyan Wang, Peng Chen, Ding Li, Chiwei Li, Qichao Zhang, Zhongpu Xia, Guizhen Yu

IEEE Robotics and Automation Letters (RA-L), 2026

Paper| video

We propose R1Sim, a tokenized traffic simulation policy that enhances human-aligned behavior generation through reinforcement learning guided by motion token entropy. By introducing entropy-based adaptive sampling and optimizing via group relative policy optimization with human-preferred rewards, R1Sim enables diverse and meaningful exploration of traffic behaviors.

ITSC

Dynamic Game-Informed Lane Changing: Leveraging Stochastic Search for Autonomous Vehicle Decision-Making

Tan Xiang^*, Ziyan Wang^*, Ding Li, Peng Chen

The IEEE International Conference on Intelligent Transportation Systems (ITSC), 2025

Paper| video

We propose Dynamic Game Informed Stochastic Search (DGSS), a novel decision-making framework that models recursive inter-vehicle interactions using multi-tree Monte Carlo Tree Search and game-theoretic principles. By simulating turn-based strategies for each surrounding vehicle and evaluating actions with a multi-objective reward, DGSS improves safety, efficiency, and interaction awareness in autonomous lane-changing.

TITS

Dynamic Origin-Destination Flow Imputation Using Feature-based Transfer Learning

Peng Chen, Ziyan Wang, Bin Zhou, Guizhen Yu

IEEE Transactions on Intelligent Transportation Systems (TITS), 2024

Paper

We developed an innovative framework utilizing an autoencoder network with feature transfer to estimate urban dynamic origin-destination flows, leveraging both connected vehicle trajectories and automatic vehicle identification data. Tested on a real-world road network, our model demonstrated superior performance, and showed robust estimation capabilities under varying observation conditions and data quality.

All Publications

Journal

Dynamic OD Estimation Using Spatio-Temporal Hybrid Graph Convolutional Network with Asynchronous Multi-Source Data, Peng Chen, Ying Guo, Ziyan Wang, Sheng Dong, Lei Wei, IEEE Transactions on Intelligent Transportation Systems (TITS), 2026
Learning Rollout from Sampling: An R1-Style Tokenized Traffic Simulation Model, Ziyan Wang, Peng Chen, Ding Li, Chiwei Li, Qichao Zhang, Zhongpu Xia, Guizhen Yu, IEEE Robotics and Automation Letters, 2026, vol. 11, no. 5, pp. 6336-6343.
Dynamic Origin-Destination Flow Imputation Using Feature-based Transfer Learning, Peng Chen, Ziyan Wang, Bin Zhou, Guizhen Yu, IEEE Transactions on Intelligent Transportation Systems, 2024, vol. 25, no. 11, pp. 17147-17159,

Conference

Dynamic Game-Informed Lane Changing: Leveraging Stochastic Search for Autonomous Vehicle Decision-Making, Tan Xiang^*, Ziyan Wang^*, Ding Li, Peng Chen, The IEEE International Conference on Intelligent Transportation Systems, 2025, pp. 4382-4388.

🎖 News

2026.03, One first-authored paper has been accepted by RA-L!
2025.07, Our team won the 3rd place in the Scenario generation track of the Third Onsite Autonomous Driving Algorithm Challenge. [Link]
2024.11, I won the National Scholarship for Graduate Students.
2024.05, Our team secured the 2nd place and innovation solution (Honorable Mention) in the track 4: Robust Depth Estimation of Robodrive Challenge. | [report] | [video]
2023.06, Our team won the 1st place in the Combined track and Intersection track of the First Onsite Autonomous Driving Algorithm Challenge. | [Link]
2022.06, Our team “A dynamic OD estimation method for urban road network based on mobile crowdsourcing and AVI data” project won the second prize in the main track of the 32nd Beihang “Feng Ru Cup”.

📖 Educations

2023.09 - 2026.07 (now), Master of Engineering, Transportation Engineering, Beihang University, Beijing, China.
2019.09 - 2023.06, Bachelor of Engineering, Transportation Engineering, Beihang University, Beijing, China.

Ziyan Wang