L4 Autonomous Driving & V2X Cooperative Perception System
Published:
Engineering validation of L4 autonomous driving, BEV sensor fusion (Camera/LiDAR/Radar), and V2X intersection coordination at MIC Munich.
Posts by Collection
portfolio
Simulation-Based Bidirectional Control between AVs and Adaptive Traffic Signal Control
Published:
Master’s Thesis & IABG-MIC Project: DQN-based traffic signal optimization with connected autonomous vehicles via SUMO/TraCI & V2X.
Toward Continually Growing World Models: An OaK-Inspired Architecture for Learned Abstraction and Planning
Published:
Structural Continual World Modeling (SCWM) operationalizing Richard Sutton’s OaK architecture and FC-STOMP lifecycle to overcome catastrophic forgetting and plasticity loss in lifelong world models.
Long-Horizon AI Task Agents & Reusable Agent Skills (SOPs)
Published:
Design and engineering of long-horizon task agents, structured agent skill workflows, and model distillation for complex task execution.
Driving World Models for Closed-Loop Simulation and Reactive Traffic Worlds
Published:
A comprehensive survey formulating a Simulation-Closure framework and claim-specific validity stack for learned driving world models.
publications
Influence of Sasobit on Physical and Chemical Properties of Asphalt
Published in Bachelor's Thesis, University of Leeds & Southwest Jiaotong University, 2023
Bachelor thesis investigating the material properties and chemical modifications of asphalt binders using Sasobit warm-mix additives.
Recommended citation: Xu, An. (2023). "Influence of Sasobit on Physical and Chemical Properties of Asphalt." Bachelor's Thesis, SWJTU-Leeds Joint School.
AI’s Transformation of Transportation in Modern Society
Published in TUM Asia Workshop on Intelligent Transportation Systems, 2025
A comprehensive study on the paradigm shift driven by AI, foundation models, and autonomous systems across modern transportation infrastructures.
Recommended citation: Xu, An. (2025). "AI's Transformation of Transportation in Modern Society." TUM Asia ITS Proceedings.
Simulation-Based Bidirectional Control between AVs and Adaptive Traffic Signal Control Using Real-World V2X-Infrastructure
Published in Master's Thesis, Technical University of Munich Asia (TUM Asia) & IABG mbH, 2026
Master thesis on bidirectional coordination between connected autonomous vehicles and adaptive signal control using SUMO, TraCI, and DQN at the IABG Mobility Innovation Campus.
Recommended citation: Xu, An. (2026). "Simulation-Based Bidirectional Control between AVs and Adaptive Traffic Signal Control Using Real-World V2X-Infrastructure." Master's Thesis, Technical University of Munich Asia.
Toward Continually Growing World Models: An OaK-Inspired Architecture for Learned Abstraction and Planning
Published in Zenodo Preprint / Workshop on Continual World Models, NeurIPS 2026, 2026
A position paper introducing Structural Continual World Modeling (SCWM), operationalizing Richard Sutton’s OaK architecture and FC-STOMP lifecycle to overcome plasticity loss and compounding error in lifelong world models.
Recommended citation: Xu, An. (2026). "Toward Continually Growing World Models: An OaK-Inspired Architecture for Learned Abstraction and Planning." Zenodo Preprint. doi:10.5281/zenodo.22103257.
Download Paper
Conditional Effects of Connected-Vehicle Requests in Reinforcement-Learning Traffic Signal Control
Published in Preprint / In Submission, 2026
A controlled factorial study disentangling queue-responsive RL adaptation from connected-vehicle request conditioning across 2,730 simulation runs and 2D crossed-bootstrap uncertainty analysis.
Recommended citation: Xu, An. (2026). "Conditional Effects of Connected-Vehicle Requests in Reinforcement-Learning Traffic Signal Control." Preprint.
Download Paper
From Foresight to Reactive Traffic Worlds: A Survey of Driving World Models for Closed-Loop Simulation
Published in SSRN Preprint / Open Research Resource, 2026
A comprehensive survey establishing a Simulation-Closure framework (Replay-Constrained, Ego-Responsive, Traffic-Reactive) and a 5-layer validity stack for learned driving world models in closed-loop autonomous driving simulation.
Recommended citation: Xu, An, Jin, Zekai, Sun, Xiaoning, Zhang, Chengbo, and Yin, Yunfei. (2026). "From Foresight to Reactive Traffic Worlds: A Survey of Driving World Models for Closed-Loop Simulation." Available at SSRN: ssrn.com/abstract=7383662 or doi.org/10.2139/ssrn.7383662.
Download Paper
talks
AI-Driven Intelligent Transportation Systems: Current State and Future Horizons
Published:
Presented research insights on the role of artificial intelligence, reinforcement learning, and world models in transforming contemporary urban traffic operations and autonomous vehicle coordination.
teaching
Teaching experience 1
Undergraduate course, University 1, Department, 2014
This is a description of a teaching experience. You can use markdown like any other post.
Teaching experience 2
Workshop, University 1, Department, 2015
This is a description of a teaching experience. You can use markdown like any other post.