Panagiotis Petropoulakis

Munich, Germany

Originally from 🇬🇷

I am Panagiotis Petropoulakis, a Ph.D. student in Computer Science at the Technical University of Munich. My research focuses on traversability, reasoning, and representation learning for mobile manipulators operating in complex and unstructured environments.

I investigate how learned state representations can encode geometric, semantic, and task-relevant properties of the environment to enable robust traversability assessment, decision-making, and motion planning. In particular, I explore methods that combine learning-based perception with structured reasoning to improve generalization and safety.

My work aims to advance autonomous operation of mobile manipulation systems in real-world domains, including construction and field robotics.

Our time is limited, so we should use it wisely. I choose to spend mine on four things I love: my passion for research, staying active and healthy through sports and nutritious food, connecting with family and friends, and volunteering. I am passionate about being involved in initiatives related to children interested in technology and women in tech.

My goal through my research and projects is to connect with people, both to offer and to learn from them, and ideally, to leave a positive impact.

For additional information, please visit my blog and have a look at my CV.

news

Aug 15, 2024	Our article State Representations as Incentives for Reinforcement Learning Agents: A Sim2Real Analysis on Robotic Grasping has been accepted at the IEEE International Conference on Systems, Man, and Cybernetics (SMC) 2024

selected publications

State Representations as Incentives for Reinforcement Learning Agents: A Sim2Real Analysis on Robotic Grasping

P. Petropoulakis^*, L. Gräf^*, M. Malmir^†, and 2 more authors

2024

Abs Bib HTML PDF Video

Choosing an appropriate representation of the environment for the underlying decision-making process of the reinforcement learning agent is not always straightforward. The state representation should be inclusive enough to allow the agent to informatively decide on its actions and disentangled enough to simplify policy training and the corresponding sim2real transfer. Given this outlook, this work examines the effect of various representations in incentivizing the agent to solve a specific robotic task: antipodal and planar object grasping. A continuum of state representations is defined, starting from hand-crafted numerical states to encoded image-based representations, with decreasing levels of induced task-specific knowledge. The effects of each representation on the ability of the agent to solve the task in simulation and the transferability of the learned policy to the real robot are examined and compared against a model-based approach with complete system knowledge. The results show that reinforcement learning agents using numerical states can perform on par with non-learning baselines. Furthermore, we find that agents using image-based representations from pre-trained environment embedding vectors perform better than end-to-end trained agents, and hypothesize that separation of representation learning from reinforcement learning can benefit sim2real transfer. Finally, we conclude that incentivizing the state representation with task-specific knowledge facilitates faster convergence for agent training and increases success rates in sim2real robot control.
@article{petropoulakis2024staterepresentationsincentives, title = {State Representations as Incentives for Reinforcement Learning Agents: A Sim2Real Analysis on Robotic Grasping}, author = {Petropoulakis, P. and Gräf, L. and Malmir, M. and Josifovski, J. and Knoll, A.}, year = {2024}, eprint = {2309.11984}, archiveprefix = {arXiv}, primaryclass = {cs.RO}, url = {https://arxiv.org/abs/2309.11984}, }