Kumar Manas joined the PEARL group at TU Darmstadt as a Postdoctoral Researcher in November 2025 under the supervision of Prof. Georgia Chalvatzaki. His current research focuses on robotic manipulation and how foundation models can be leveraged for generalizable robot learning — enabling agents to reason, act, and adapt across diverse manipulation tasks.
Before joining PEARL, he worked on his Ph.D. in Mathematics and Computer Science at Freie Universität Berlin, where his research focused on automating the formalization of natural language specifications in formal logic using language models and integrating these representations into robotic systems. He was supervised by Prof. Dr. Adrian Paschke. During his Ph.D., he also worked as a Research Scientist at Continental’s AI Lab (Berlin), developing generative model-based planners and uncertainty-aware trajectory prediction systems for automated driving.
His research bridges robot learning, generative modeling, and multimodal foundation models, aiming to build intelligent robotic systems that combine structured reasoning, perception, and control for safe and flexible interaction with the world.
Research Interests
* Memory design for VLA and World models
* Foundation Models for Robotics
* Robot Learning and Manipulation
* Generative and Diffusion-Based Planning
* Knowledge-Guided and Uncertainty-Aware Decision Making
Selected Publications
* Uncertainty-Aware Trajectory Prediction via Rule-Regularized Heteroscedastic Deep Classification, K Manas, C Schlauch, A Paschke, C Wirth, N Klein, Robotics: Science and Systems (RSS) 2025.
* Knowledge Integration Strategies in Autonomous Vehicle Prediction and Planning: A Comprehensive Survey, K Manas, A Paschke, IEEE Intelligent Vehicles Symposium (IV), 2025.
* CoT-TL: Low-Resource Temporal Knowledge Representation of Planning Instructions Using Chain-of-Thought Reasoning, K Manas, S Zwicklbauer, A Paschke, IEEE/RSJ IROS 2024.
* TR2MTL: LLM-based Framework for Metric Temporal Logic Formalization of Traffic Rules , K Manas, S Zwicklbauer, A Paschke, IEEE Intelligent Vehicles Symposium (IV) 2024.
* Semantic Role Assisted Natural Language Rule Formalization for Intelligent Vehicle , K Manas, A Paschke, International Joint Conference on Rules and Reasoning (RuleML+RR) 2023.
*(Full list available on Google Scholar)*
Contact
kumar.manas<at>tu-darmstadt.de
TU Darmstadt, Landwehrstr. 50A, 64293 Darmstadt
Office: Room 01, Building S4|23
