Zewen Yang
AI Researcher | Robotic Control Engineer
I am a Senior Research Scientist at Agile Robots in Munich, working on Vision-Language-Action models, World Action Models, generative motion-planning methods, and agentic AI frameworks that enable robots to perceive, reason, plan, and act. My mission is to turn advances in machine learning into safe, reliable systems that solve real-world problems.
Previously, I was a postdoctoral researcher at the Munich Institute of Robotics and Machine Intelligence (MIRMI), supervised by Prof. Sami Haddadin and Dr. Hamid Sadeghian. Before that, in the aftermath of the pandemic, I joined the Robert Koch Institute as a postdoctoral researcher, working on AI for public health research. I did my Ph.D. research in Control Science & Engineering at the Technical University of Munich and Harbin Engineering University, where I was supervised by Prof. Sandra Hirche, Prof. Zheping Yan, and Dr. Stefan Sosnowski.
My work lies at the intersection of Machine Learning, Control, and Robotics, driven by a strong interest in machine intelligence and the emerging technologies that make everyday life better and work more efficient. I enjoy moving between building methods that hold up in the real world and exploring the deeper scientific questions behind them. I also care deeply about education and research, and find it especially rewarding to work alongside passionate people and ambitious students, growing together with them.
I am always glad to connect with others who share these interests, whether for collaboration, projects, research, or simply a good conversation. Feel free to drop me a line at zewenreal [at] gmail [dot] com.
News
| Aug 29, 2026 | I co-organized From Kinematics to Motion: Hands-On Robotics with reBot Arm in Garching, the second OpenELAB × Seeed Studio workshop, taking robot kinematics and control from concepts onto the real arm. 🤖 (announcement) |
|---|---|
| Aug 19, 2026 | Our paper UniConFlow: A Unified Constrained Flow-Matching Framework for Certified Motion Planning has been accepted to IEEE Transactions on Robotics (T-RO)! 🎉 |
| Aug 05, 2026 | Our paper Whom to Trust? Selective Online Learning in Multi-Agent Systems with Prior-Aware Gaussian Process Regression, the extended journal version of our AAMAS conference paper, has been accepted to JAAMAS! ✨ |
| Aug 01, 2026 | I co-organized the Physical AI Workshop: Hands-on reBot Arm in Garching-Hochbrück, Munich, a full-day session co-hosted by OpenELAB and Seeed Studio, covering assembly, calibration, data collection, and model training on real hardware. 🤖 (announcement) |
| Jun 17, 2026 | Two of our papers have been accepted to IROS 2026! 🚀 |
Blog
Selected publications
- JAAMASWhom to Trust? Selective Online Learning in Multi-Agent Systems with Prior-Aware Gaussian Process RegressionAutonomous Agents and Multi-Agent Systems, 2026Accepted. Impact factor: 2.4, JCR: Q2
- IROSSafe-Night VLA: Seeing the Unseen via Thermal-Perceptive Vision-Language-Action Models for Safety-Critical ManipulationIn Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems, 2026Oral presentation (acceptance rate 36%)
- ICRARCM Constraint-Consistent Dynamic Control in Surgical RobotsIn Proceedings of the IEEE International Conference on Robotics and Automation, 2026Oral presentation (top 1% in submissions)
- ICRAContact-Safe Reinforcement Learning with ProMP Reparameterization and Energy AwarenessIn IEEE International Conference on Robotics and Automation (ICRA), 2026Poster presentation (acceptance rate 38.04%)
- IEEE JSTSPMachine Learning based Controller using Gaussian Process Regression in Electric Motors: Interpretability, Safety and ReliabilityIEEE Journal of Selected Topics in Signal Processing, 2026Impact factor: 13.7, JCR: Q1
- EAAISafe event-triggered control of unmanned surface vehicles with Gaussian processes: Resilience in denial of service attacks and uncertain dynamicsEngineering Applications of Artificial Intelligence, 2025Impact factor: 9.0, JCR: Q1
- IEEE TNNLSCooperative Online Learning for Multiagent System Control via Gaussian Processes With Event-Triggered MechanismIEEE Transactions on Neural Networks and Learning Systems, 2025Impact factor: 9.7, JCR: Q1
- IEEE/ASME TMECHTorque-Induced-Overshoot Reduction Inspired Compensator for PMSMs Using Motor-Physics Embedded Gaussian Process RegressionIEEE/ASME Transactions on Mechatronics, 2025Impact factor: 6.3, JCR: Q1
- ACCCooperative Learning with Gaussian Processes for Euler-Lagrange Systems Tracking Control Under Switching TopologiesIn 2024 American Control Conference (ACC), 2024Oral presentation
- CCCKernel-based Learning for Safe Control of Discrete-Time Unknown Systems under Incomplete ObservationsIn 2024 43rd Chinese Control Conference (CCC), 2024Oral presentation
- EJCDecentralized event-triggered online learning for safe consensus control of multi-agent systems with Gaussian process regressionEuropean Journal of Control, 2024Impact Factor: 2.6, JCR: Q2
- IEEE TACCooperative Control of Uncertain Multiagent Systems via Distributed Gaussian ProcessesIEEE Transactions on Automatic Control, 2023Impact factor: 7.0, JCR: Q1
- L4DCCan Learning Deteriorate Control? Analyzing Computational Delays in Gaussian Process-Based Event-Triggered Online LearningIn Proceedings of The 5th Annual Learning for Dynamics and Control Conference, 15–16 jun 2023Poster presentation
- IECONLearning-based Control for PMSM Using Distributed Gaussian Processes with Optimal Aggregation StrategyIn IECON 2023- 49th Annual Conference of the IEEE Industrial Electronics Society, 2023Oral presentation
- CDCDistributed Learning Consensus Control for Unknown Nonlinear Multi-Agent Systems based on Gaussian ProcessesIn 2021 60th IEEE Conference on Decision and Control (CDC), 2021Oral presentation
- OEVirtual leader based path tracking control for Multi-UUV considering sampled-data delays and packet lossesOcean Engineering, 2020Impact factor: 6.3, JCR: Q1
- OEDiscrete-time coordinated control of leader-following multiple AUVs under switching topologies and communication delaysOcean Engineering, 2019Impact factor: 6.3, JCR: Q1
- OCEANSCoordinated Control for Trajectory Tracking of Multiple UUVs with Input SaturationIn OCEANS 2019 - Marseille, 2019Oral presentation
- CCCDiving Control of Underactuated UUV Based on Backstepping Upper Bound Sliding Mode MethodIn 2019 Chinese Control Conference (CCC), 2019Oral presentation
- IECONDiscrete-time Path Tracking Control of Multiple UUVs Based on Virtual Leader under Time Varying DelayIn IECON 2018 - 44th Annual Conference of the IEEE Industrial Electronics Society, 2018Oral presentation
- CCCHorizontal Trajectory Tracking Control of AUV Using a Two-way Channel High Gain ObserverIn 2018 37th Chinese Control Conference (CCC), 2018Oral presentation