Papers
20
Total Citations
774
H-Index
14
About
Sanjay Krishnan is a leading researcher at the intersection of robot learning, surgical robotics, and autonomous systems. His work focuses on enabling robots to learn complex manipulation tasks from human demonstrations, with a particular emphasis on the challenging domain of robot-assisted surgery. Krishnan pioneered the use of unsupervised trajectory segmentation, introducing Transition State Clustering (TSC) and its deep learning variant TSC-DL, which automatically decompose surgical demonstrations into meaningful, reusable skills—a foundational contribution cited over 160 times. He also developed SWIRL, a hybrid inverse reinforcement learning algorithm that combines expert demonstrations with autonomous exploration for tasks with delayed rewards. In surgical robotics, Krishnan advanced autonomous debridement with cable-driven robots, proposing a two-phase calibration procedure to overcome kinematic nonlinearities, and tackled automated camera control for the da Vinci Research Kit. His work on deep continuous options (DDCO) and multi-level option discovery has further pushed the boundaries of hierarchical reinforcement learning. With over 600 total citations across his most-cited papers, Krishnan’s research has significantly shaped how robots learn from limited, imperfect demonstrations, directly impacting the safety and efficiency of autonomous surgical systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5
- 6Multi-Level Discovery of Deep Options70 citations · 2017
- 7
- 8
- 9
- 10Learning 2D Surgical Camera Motion From Demonstrations30 citations · 2018