Dhruv Metha RameshI am a PhD student at the Robot Learning Lab at Rutgers University - New Brunswick. My research is advised by Dr. Abdeslam Boularias and Dr. Kostas E. Bekris. I'm interested in Long-Horizon Task and Motion Planning and Scene Estimation for Robotics. Email / GitHub / Google Scholar / CV |
|
Research |
PROBE: Proprioceptive Obstacle Detection and Estimation while Navigating in ClutterDhruv Metha Ramesh, Shreesh Keskar, Aravind Sivaramakrishnan, Kostas E Bekris, Jingjin Yu, Abdeslam Boularias Under Review, 2024 paper / code / website / PROBE introduces a transformer based system that predicts 2D planar obstacles in the environment only from the robot’s interaction with the obstacles. The transformer uses only the history of the robot’s proprioceptive states (joint positions, joint velocities, torques applied on the joints and robot pose) for this prediction. |
|
Integrating Model-based Control and RL for Sim2Real Transfer of Tight Insertion PoliciesIsidoros Marougkas*, Dhruv Metha Ramesh*, Joe H. Doerr, Edgar Granados, Aravind Sivaramakrishnan, Abdeslam Boularias, Kostas E. Bekris Under Review, 2024 paper / code / website / This work addresses precise (<1mm) object insertion tasks by combining model-based control with residual RL, enabling zero-shot transfer from simulation to reality without fine-tuning. The integrated policy outperforms state-of-the-art RL and hybrid approaches, validated through simulations and real-world tests. |
Reviewing Service |
|
|
Design and source code from Jon Barron's website |