
Gokul Puthumanaillam
I am a PhD student at the University of Illinois Urbana-Champaign (UIUC) where I am advised by Prof. Melkior Ornik (Learning, Autonomy, Decision, and Control (LEADCAT) Lab). I finished my bachelors in ECE+Math.
Research
My research is on robot learning under sparsity of data and scarcity of resources: building robot foundation models that learn from limited interaction and remain deployable on real hardware. My research threads:
(i) Pretraining objectives and data generation for policies that must learn from scarce or weakly supervised interaction.
(ii) Theory and methods for inference-time efficiency, making generative policies viable for real-time control.
(iii) Formal guarantees for learned controllers operating under unknown dynamics.
All of my papers are evaluated in real-world, real-time, in-the-wild robotic settings on hardware, and are theoretically grounded in machine learning (Reinforcement Learning, Diffusion Models, Vision-Language Action Models, World Models, World Action Models, Robot Vision etc.).
At the systems level, I focus on implementation and speedups for ML algorithms and robotic systems--building end-to-end pipelines, optimizing bottlenecks, and pushing planning and learning to real-time on hardware. (I regularly cry while implementing neural networks in C++). I am passionate about developing software for hardware, open-source contributions and translating complex theories into practical applications.
Hopefully, there will be time for everything.
★ RSS 2026 Muninn won the Outstanding Student Paper Award and was also a Finalist for the Outstanding Paper Award at RSS 2026!
I am excited to be spending the summer of 2026 with (Seattle Robotics Lab) as a robotics research intern.
I spent the summer of 2025 at Research as a research scientist intern in the multimodal group.
Publications
- [RSS 2026] ★ Outstanding Student Paper ★ Finalist: Outstanding Paper Muninn: Your Trajectory Diffusion Model But Faster.
@article{puthumanaillam2026muninn,
title={Muninn: Your Trajectory Diffusion Model But Faster},
author={Gokul Puthumanaillam and Hao Jiang and Ruben Hernandez and Jose Fuentes and Paulo Padrao and Leonardo Bobadilla and Melkior Ornik},
journal={{arXiv preprint arXiv:2605.09999},
year={2026}
}
- [CoRL 2026] Trajectory-Level Redirection Attacks on Vision-Language-Action Models.
@misc{puthumanaillam2026trajectorylevelredirectionattacksvisionlanguageaction,
title={Trajectory-Level Redirection Attacks on Vision-Language-Action Models},
author={Gokul Puthumanaillam and Vardhan Dongre and Pranay Thangeda and Hooshang Nayyeri and Dilek Hakkani-Tür and Melkior Ornik},
year={2026},
eprint={2606.12978},
archivePrefix={arXiv},
primaryClass={cs.RO},
url={https://arxiv.org/abs/2606.12978},
note={Accepted at the Conference on Robot Learning (CoRL 2026)}
}
- [IROS 2026] ★ Long talk Amortizing Trajectory Diffusion with Keyed Drift Fields.
@article{puthumanaillam2026amortizing,
title={Amortizing Trajectory Diffusion with Keyed Drift Fields},
author={Puthumanaillam, Gokul and Ornik, Melkior},
journal={arXiv preprint arXiv:2603.14056},
year={2026}
}
- [CoRL 2026] When Should a Robot Replan? Regret-Guided Update Scheduling in Time-Varying MDPs.
@misc{musavi2026robotreplanregretguidedupdate,
title={When Should a Robot Replan? Regret-Guided Update Scheduling in Time-Varying {MDPs}},
author={Negin Musavi and Gokul Puthumanaillam and Ruben Hernandez and William Schafer and Melkior Ornik},
year={2026},
eprint={2606.16972},
archivePrefix={arXiv},
primaryClass={cs.RO},
url={https://arxiv.org/abs/2606.16972},
note={Accepted at the Conference on Robot Learning (CoRL 2026)}
}
- [CoRL 2026] Task-Aware Environment Augmentation for Reliable Navigation via Shielded Conditional Diffusion.
@misc{phoompho2026taskawareenvironmentaugmentation,
title={Task-Aware Environment Augmentation for Reliable Navigation via Shielded Conditional Diffusion},
author={Bharawee Phoompho and Gokul Puthumanaillam and Yan Miao and Ruben Hernandez and Tim Bretl and Sayan Mitra and Melkior Ornik},
year={2026},
eprint={2606.15154},
archivePrefix={arXiv},
primaryClass={cs.RO},
url={https://arxiv.org/abs/2606.15154},
note={Accepted at the Conference on Robot Learning (CoRL 2026)}
}
- [RLC 2026] SCoUT: Scalable Communication via Utility-Guided Temporal Grouping in Multi-Agent Reinforcement Learning.
@article{vora2026scout,
title={SCoUT: Scalable Communication via Utility-Guided Temporal Grouping
in Multi-Agent Reinforcement Learning},
author={Vora, Manav and Puthumanaillam, Gokul and Tsukamoto, Hiroyasu
and Ornik, Melkior},
journal={arXiv preprint arXiv:2603.04833},
year={2026}
}
- [CoRL 2025] ★ Oral (Top 5%) Belief-Conditioned One-Step Diffusion: Real-Time Trajectory Planning with Just-Enough Sensing.
@article{bcod2025,
title = {Belief-Conditioned One-Step Diffusion: Real-Time Trajectory Planning with Just-Enough Sensing},
author = {Gokul Puthumanaillam and Aditya Penumarti and Manav Vora and Paulo Padrao and Jose Fuentes and Leonardo Bobadilla and Jane Shin and Melkior Ornik},
year = {2025},
eprint = {2508.12166},
archivePrefix = {arXiv},
primaryClass = {cs.RO},
url = {https://arxiv.org/abs/2508.12166}
}
- [IROS 2026] Online Learning of Deceptive Policies under Intermittent Observation.
@misc{puthumanaillam2025onlinelearningdeceptivepolicies,
title={Online Learning of Deceptive Policies under Intermittent Observation},
author={Gokul Puthumanaillam and Ram Padmanabhan and Jose Fuentes and Nicole Cruz and Paulo Padrao and Ruben Hernandez and Hao Jiang and William Schafer and Leonardo Bobadilla and Melkior Ornik},
year={2025},
eprint={2509.14453},
archivePrefix={arXiv},
primaryClass={cs.RO},
url={https://arxiv.org/abs/2509.14453},
}
- [RA-L 2026] GUIDEd Agents: Enhancing Navigation Policies through Task-Specific Uncertainty Abstraction in Localization-Limited Environments
@article{puthumanaillam2024enhancing,
title={GUIDEd Agents: Enhancing Navigation Policies through Task-Specific Uncertainty
Abstraction in Localization-Limited Environments},
author={Gokul Puthumanaillam and Paulo Padrao and Jose Fuentes
and Leonardo Bobadilla and Melkior Ornik},
journal={IEEE Robotics and Automation Letters},
year={to be published. for now arXiv preprint arXiv:2410.15178},
publisher={IEEE}
}
- [IROS 2025] ★ Oral TRACE: A Self-Improving Framework for Robot Behavior Forecasting with Vision-Language Models.
@article{trace2025,
title = {TRACE: A Self-Improving Framework for Robot Behavior Forecasting with Vision-
Language Models},
author = {Gokul Puthumanaillam and Paulo Padrao and
Jose Fuentes and Pranay Thangeda and William E. Schafer
and Jae Hyuk Song and Karan Jagdale and Leonardo Bobadilla
and Melkior Ornik},
year = {2025},
eprint = {2503.00761},
archivePrefix = {arXiv},
primaryClass = {cs.RO},
url = {https://arxiv.org/abs/2503.00761}
}
- [IROS 2025] ★ Oral Motion Planning and Control with Unknown Nonlinear Dynamics through Predicted Reachability.
@article{zhang2025motion,
title={Motion Planning and Control with Unknown Nonlinear Dynamics through
Predicted Reachability},
author={Zhang, Zhiquan and Puthumanaillam, Gokul and Vora, Manav
and Ornik, Melkior},
journal={arXiv preprint arXiv:2503.03633},
year={2025}
}
- [AAMAS 2025] Enhancing Robot Navigation Policies with Task-Specific Uncertainty Management.
@article{puthumanaillam2024enhancing,
title={Enhancing Robot Navigation Policies with Task-Specific Uncertainty Management},
author={Gokul Puthumanaillam and Paulo Padrao and Jose Fuentes
and Leonardo Bobadilla and Melkior Ornik},
journal={arXiv preprint arXiv:2410.15178},
year={2024},
note={Accepted at International Conference on Autonomous Agents and Multiagent Systems
(AAMAS 2025)}
}
- [L4DC 2025] TAB-Fields: A Maximum Entropy Framework for Mission-Aware Adversarial Planning.
@article{puthumanaillam2024tabfields,
title={TAB-Fields: A Maximum Entropy Framework for Mission-Aware Adversarial Planning},
author={Gokul Puthumanaillam and Jae Hyuk Song and Nurzhan Yesmagambet and
Shinkyu Park and Melkior Ornik},
journal={arXiv preprint arXiv:2412.02570},
year={2024},
note={Accepted for presentation at L4DC 2025}
}
- [ACE 2025] The Lazy Student's Dream: ChatGPT Passing an Engineering Course on Its Own.
@article{puthumanaillam2025lazy,
title={The Lazy Student's Dream: ChatGPT Passing an Engineering Course on Its Own},
author={Puthumanaillam, Gokul and Bretl, Thomas and Ornik, Melkior},
journal={arXiv preprint arXiv:2503.05760},
year={2025}
}
- [IROS 2024] ★ Oral, Top 10% ComTraQ-MPC: Meta-Trained DQN-MPC Integration for Trajectory Tracking with Limited Active Localization Updates.
@article{puthumanaillam2024comtraq,
title={ComTraQ-MPC: Meta-Trained DQN-MPC Integration for Trajectory Tracking with
Limited Active Localization Updates},
author={Puthumanaillam, Gokul and Vora, Manav and Ornik, Melkior},
journal={arXiv preprint arXiv:2403.01564},
year={2024},
note={Accepted at IEEE/RSJ International Conference on Intelligent Robots and Systems
(IROS 2024)}
}
- [ICRA 2024] Weathering ongoing uncertainty: learning and planning in a time-varying partially observable environment.
@article{puthumanaillam2023weathering,
title={Weathering ongoing uncertainty: learning and planning in a
time-varying partially observable environment},
author={Puthumanaillam, Gokul and Liu, Xiangyu
and Mehr, Negar and Ornik, Melkior},
journal={arXiv preprint arXiv:2312.03263},
year={2023},
note={accepted at IEEE International Conference on Robotics and Automation
(ICRA) 2024}
}
- [Neurocomputing 2023] ★ Best Paper Award at IEEE GRSS Texture based prototypical network for few-shot semantic segmentation of forest cover: generalizing for different geographical regions.
@article{puthumanaillam2023texture,
title={Texture based prototypical network for few-shot semantic segmentation of
forest cover: Generalizing for different geographical regions},
author={Puthumanaillam, Gokul and Verma, Ujjwal},
journal={Neurocomputing},
volume={538},
pages={126201},
year={2023},
publisher={Elsevier},
note={Best Paper Award at IEEE GRSS}
}
Earlier work 7 papers
- [AIAA SciTech 2025] Modeling and Analysis of Collision Avoidance with ACAS Xu in Verse for the GUAM Lift+ Cruise Model.
@inproceedings{bullock2025modeling,
title={Modeling and Analysis of Collision Avoidance with ACAS Xu in Verse for the GUAM Lift+ Cruise Model},
author={Bullock, JL and Song, L and Puthumanaillam, G and Li, Y and Cheng, S and Ornik, M and Hovakimyan, N},
booktitle={AIAA SCITECH 2025 Forum},
pages={1323},
year={2025}
}
- [AIAA SciTech 2024] Assured Collision Avoidance for Learned Controllers: A Case Study of ACAS Xu.
@inproceedings{puthumanaillam2024assured,
title={Assured Collision Avoidance for Learned Controllers: A Case Study
of ACAS Xu},
author={Puthumanaillam, Gokul and Vora, Manav Ketan and Shafa,
Taha and Li, Yangge and Ornik, Melkior and Mitra, Sayan},
booktitle={AIAA SCITECH 2024 Forum},
pages={1168},
year={2024}
}
- [AIAA SciTech 2024] Online Learning and Planning in Time-Varying Environments: An Aircraft Case Study.
@inproceedings{puthumanaillam2024online,
title={Online Learning and Planning in Time-Varying Environments: An Aircraft Case Study},
author={Puthumanaillam, Gokul and Mamik, Yuvraj and Ornik, Melkior},
booktitle={AIAA SCITECH 2024 Forum},
pages={0109},
year={2024}
}
- [ICUAS 2024] A dynamic area prioritization based adaptive multi-robot target search.
@inproceedings{gokul2024dynamic,
title={A Dynamic Area Approximation-Based Stochastic Multi-UAV Target Search with Noisy
Measurements},
author={Gokul, P and Harikumar, K and Senthilnath, J},
booktitle={2024 International Conference on Unmanned Aircraft Systems (ICUAS)},
pages={718--723},
year={2024},
organization={IEEE}
}
- [IGARSS 2023] Improved semantic segmentation for identification of flooded regions in UAV aerial images: a transformer-based approach.
@inproceedings{verma2023improved,
title={Improved Semantic Segmentation for Identification of Flooded Regions
in UAV Aerial Images: A Transformer-Based Approach},
author={Verma, Ujjwal and Puthumanaillam, Gokul},
booktitle={IGARSS 2023-2023 IEEE International Geoscience and Remote Sensing Symposium},
pages={4800--4803},
year={2023},
organization={IEEE}
}
- [FABS 2021] Pandora: an IOT based intrusion detection honeypot with real-time monitoring.
@inproceedings{ramakrishnan2021pandora,
title={Pandora: An IOT based Intrusion Detection Honeypot with Real-time Monitoring},
author={Ramakrishnan, Karthik and Gokul, P and Nigam, Rohan},
booktitle={2021 International Conference on Forensics, Analytics, Big Data, Security (FABS)},
volume={1},
pages={1--7},
year={2021},
organization={IEEE}
}
- [IROS 2021 Workshop] BlockMapping: A Blockchain based multi-robot exploration.
@inproceedings{puthumanaillam2022blockmapping,
author = {Puthumanaillam, Gokul and Paramjit, Singh and Harshwardhan, Kulkarani},
title = {BlockMapping: A Blockchain based multi-robot exploration},
booktitle = {Blockchain Applications in Robotics and Automation, IROS 2021},
year = {2021}
}
Theses
- Zero-shot segmentation on the edge. Bachelors (Hons) Thesis. Advisor: Prof. Ujjwal Verma. 2022.
- SOLO Design Report (IGVC 2019). Available at [PDF]. 2019.
- MBZIRC 2023 White paper: Team Luna. Available at: [PDF]. 2019.
- Zeoco: An insight into daily carbon footprint consumption. 2021. arXiv: 2102.06185
Patent Applications
- "POMCP based MIC creation for engines". Patent Application No. 606950-US-2 (146674.582103). Patent application under review. 2022 with GE Aerospace.
- "Estimation of environmental disturbance for trailering". Patent Application No. ENT-2023-023. Patent application under review. 2023 with Brunswick Corporation.
Recent News
- Sep'26: 3× papers accepted at CoRL 2026!
- Jul'26: Muninn won the Outstanding Student Paper Award at RSS 2026!!!
- Jul'26: Muninn was nominated for the Outstanding Paper Award and the Outstanding Student Paper Award at RSS 2026
- Jun'26: 2× papers accepted at IROS 2026!
- May'26: Muninn accepted @ RSS 2026
- May'26: SCoUT accepted @ RLC 2026
- May'26: Started my internship at NVIDIA Research
- Sep'25: GUIDEd Agents accepted @ RA-L
- Aug'25: B-COD accepted @ CoRL 2025 oral (top 5%)
- Jun'25: Awarded the Young Author Award by IFAC
- Jun'25: 2× papers accepted at IROS 2025!
- TRACE · oral
- Predicted Reachability · oral
- May'25: Started my internship at Adobe Research
- Feb'25: TAB-Fields accepted @ L4DC 2025
- Dec'24: Paper accepted @ AAMAS 2025
- Jul'24: ComTraQ-MPC accepted @ IROS 2024 oral (top 10%)
- Mar'24: Received Department Fellowship
- Mar'24: Received travel grant for ICRA 2024
- Jan'24: Cleared PhD Qualifying Exam
- Jan'24: Planning in TV-POMDPs accepted @ ICRA 2024
- Apr'24: Paper accepted at ICUAS 2024
- Dec'23: NNCS Verification accepted @ AIAA SciTech 2024
- Dec'23: Application of TV-POMDP to GUAM accepted @ AIAA-SIAM 2024 also presented at ULI
- Jun'23: Best Paper Award @ IEEE GRSS 2023
- Jun'23: Texture guided PANet accepted @ Neurocomputing
- May'23: Paper accepted @ IGARSS 2023
Scroll for older updates ↓
Open Source
I am passionate about open-source software. Reference implementations of all algorithms I designed at LEADCAT are publicly available. I have also contributed features and bug-fixes to projects like Gazebo, Webots, Ardupilot, and Arduino.
- Released: Pathy: Filesystem path completions inside Python string literals for Zed. (v1.0.0 released: 01/02/2026 - Find it here)
- Released: Belief-Conditioned One-Step Diffusion (v1.0.0 released: 05/10/2025 - Find it here)
- Released: BlueROV2 Gymnasium Environment (v2.0.0 released: 11/11/2024 - Find it here)
- Released: Code for ComTraQ-MPC (IROS 2024) (v1.0.0 released: 08/19/2024 - Find it here)
- Released: ROS2 drivers for u-blox ZED F9P GPS receivers (v2.0.2 released: 07/10/2023 - Find it here)
- Upcoming: Vim plugin for ROS2 developers (Expected: Spring, 2026)
- Upcoming: ROS2 based space robotics with time-varying environments (Expected: Spring, 2026)
Academic Service
Conferences: CVPR (2026), ICRA (2026, 2025, 2024), IROS (2025, 2024), L4DC (2026, 2025), ICLR (2026, 2025), CoRL (2025), CDC (2024), ITSC (2025), AIAA SciTech (2024), AICECS(Student chair; 2022)
Journals: RAL, Nature Communications, OJCSYS
Acknowledgments
It takes a village to raise a child.
My parents have made many sacrifices to support me. My sister has always inspired and encouraged me. My advisors gave me the right environment and guidance to grow. My friends are the ones who've believed in me at my lowest, and celebrated with me at my highest. If I have seen further, it is by standing on shoulders of these giants.
For who I am and what I have to offer, I owe it ALL to them.