Subramanya Nageshrao

About

Subramanya Nageshrao is a leading researcher at the intersection of reinforcement learning (RL) and robotics, with a primary focus on developing safe, interpretable, and high-performance control systems for autonomous agents. His major contributions lie in advancing actor-critic RL methods for continuous control, particularly by integrating them with classical control theory to enhance robustness and tracking accuracy. His most cited work, "Reinforcement learning based compensation methods for robot manipulators" (2018, 122 citations), demonstrates how RL can augment nominal feedback controllers to significantly improve precision in robotic manipulation tasks. Nageshrao has also pioneered approaches to make RL policies more transparent, as seen in his recent work on "Interpretable Reinforcement Learning for Robotics and Continuous Control" (2023), addressing a critical barrier to deploying learned policies in safety-critical domains like autonomous driving and industrial robotics. His research on "Learning Complex Behaviors via Sequential Composition and Passivity-Based Control" (2015) further showcases his ability to combine learning with formal guarantees of stability. With a career spanning both theoretical foundations and experimental validation, Nageshrao’s work is essential reading for anyone interested in bridging the gap between modern deep RL and practical, reliable robotic systems.

Research Focus

Key Achievements

Top Papers

1. 1
   Reinforcement learning based compensation methods for robot manipulators

   122 citations · 2018
2. 2
   Actor-critic reinforcement learning for tracking control in robotics

   30 citations · 2016
3. 3
   Learning Complex Behaviors via Sequential Composition and Passivity-Based Control

   15 citations · 2015
4. 4
   Interpretable Reinforcement Learning for Robotics and Continuous Control

   2 citations · 2023

Key Collaborators

Contact & Links