Xiao Wang, M.Sc.
Technische Universität München
Informatik 6 - Lehrstuhl für Robotik, Künstliche Intelligenz und Echtzeitsysteme (Prof. Knoll)
Postadresse
Postal:
Boltzmannstr. 3
85748 Garching b. München
- Tel.: +49 (89) 289 - 18111
- Sprechstunde: Appointment by email
- Raum: 5607.03.042
- xiao.wang@tum.de
Curriculum Vitae
Xiao Wang is a research assistant and PhD student in the Cyber-Physical Systems Group under Prof. Dr.-Ing. Matthias Althoff since 2019. She graduated with the Master of Science degree in Mechanical Engineering from the Technische Universität München, Germany, in 2018. She received the Bachelor of Engineering degree in Vehicle Engineering from Tongji University, Shanghai, China.
Her research focuses on motion planning for autonomous vehicles, formal methods, and safe reinforcement learning.
Offered Theses:
Open theses:
Currently, there are no open topics for theses.
If you are interested in this topic or in her research and want to write a thesis in this area, please feel free to contact her. General ideas could be to apply formal method to standard RL techniques to increase safety, to benchmark state-of-art safe RL algorithms, to learn driving behaviors from real traffic data, or to increase sample efficiency of RL for motion planning, etc.
Supervised theses:
- [BT | SS19]: Christoph Pillmayer - "Online Verification for Autonomous Vehicles using Motion Primitives and Deep Reinforcement Learning"
- [BT | SS19]: Hagen Winkelmann - "Learning Cost Functions for Sampling Based Planners in Autonomous Driving"
- [MT | WS19]: Hanna Krasowski - "Safe Reinforcement Learning for Autonomous Vehicles"
- [MT | SS20]: Xi Chen - "Learning Driving Policies Using Reinforcement Learning Combined with Sampling-based Motion Planner"
- [MT | SS20]: Kailiang Dong - "Generative Adversarial Imitation Learning for Highway Autonomous Driving with Safety Guarantees"
- [MT | WS20]: Zhenyu Li - "Safe Reinforcement Learning for Continuous Control Tasks"
- [MT | SS21]: Jiaying Huang - "An Online Verification Framework for Autonomous Driving Using Sampling-Based Motion Planners"
Ongoing theses:
- MT: Daniel Tar - "Ensuring Safety for Reinforcement-Learning-Based Motion Planners Using Online Verification"
- MT: Guyue Huang - "Safety Falsification for Black-Box Motion Planners of Autonomous Vehicles"
- MT: Jinyue Guan - "Ensuring Drivability of Fail-safe Trajectories for Autonomous Vehicles using Reinforcement Learning Methods"
- MT: Christoph Pillmayer - "Constrained Reinforcement Learning for Autonomous Driving"
- MT: Murat Can Üste - "Generation of Naturalistic Traffic Rule Violations Using Imitation Learning"
Teaching
WS 2020/21
- Exercise: Techniques in Artificial Intelligence
- Practical Course: Motion Planning for Autonomous Vehicles
SS 22
- Seminar: Cyber-Physical Systems
- Practical Course: Motion Planning for Autonomous Vehicles
WS 2019/20
- Exercise: Techniques in Artificial Intelligence
SS 19
- Seminar: Cyber-Physical Systems
- Master Practical Course: Motion Planning for Autonomous Vehicles
WS 2018/19
- Exercise: Techniques in Artificial Intelligence
- Practical Course: Motion Planning for Autonomous Vehicles
Publications
2021
- CommonRoad-RL: A Configurable Reinforcement Learning Environment for Motion Planning of Autonomous Vehicles. IEEE International Conference on Intelligent Transportation Systems (ITSC), 2021 more… BibTeX Full text ( DOI ) Full text (mediaTUM)
2020
- Trajectory Prediction for Intelligent Vehicles Using Spatial-Attention Mechanism. IET Intelligent Transport Systems, 2020 more… BibTeX Full text ( DOI )
- Safe Reinforcement Learning for Autonomous Lane Changing Using Set-Based Prediction. 2020 IEEE International Conference on Intelligent Transportation Systems (ITSC), 2020 more… BibTeX Full text ( DOI ) Full text (mediaTUM)
- Coupling Apollo with the CommonRoad Motion Planning Framework. FISITA World Congress, 2020 more… BibTeX Full text (mediaTUM)
- Falsification-Based Robust Adversarial Reinforcement Learning. Proc. of IEEE International Conference on Machine Learning and Applications, 2020 more… BibTeX Full text ( DOI ) Full text (mediaTUM)