Rohan Paul

About

Rohan Paul is a versatile robotics researcher whose work spans mobile robotics, robot manipulation, natural language understanding, and autonomous perception. His career bridges foundational hardware challenges and cutting-edge AI-driven robot cognition, making him a significant figure across multiple generations of robotics research. Paul's early contributions tackled critical mechanical and control challenges in robot manipulators, including pioneering joint torque control through direct feedback (170 citations), analyzing wrist force sensor stiffness effects, and developing hybrid position-force control strategies — work that remains influential in manipulation research. His study of wrist singularities further demonstrated a sustained commitment to robust manipulator design. In mobile robotics, Paul co-created the widely adopted New College Vision and Laser Dataset (310 citations), providing the community with a rich benchmark for outdoor autonomous navigation. His work on shadow-invariant image representations for visual localization addressed a practical yet underappreciated challenge in outdoor robot perception. More recently, Paul has championed the grounding of natural language instructions for robots, developing efficient models that map abstract spatial language to real-world robot actions across manipulator and mobile platforms (90 and 59 citations respectively). His research on introspective classification and multimodal semantic knowledge further reflects a commitment to building robots that reason reliably under uncertainty — an increasingly essential capability as robots move into unstructured human environments.

Research Focus

Key Achievements

Top Papers

1. 1
   The New College Vision and Laser Data Set

   310 citations · 2009
2. 2
   Joint torque control by a direct feedback for industrial robots

   170 citations · 1983
3. 3
   Dealing with shadows: Capturing intrinsic scene appearance for image-based outdoor localisation

   95 citations · 2013
4. 4
   Efficient Grounding of Abstract Spatial Concepts for Natural Language Interaction with Robot Manipulators

   90 citations · 2016
5. 5
   The effect of wrist force sensor stiffness on the control of robot manipulators

   90 citations · 2005
6. 6
   Hybrid control of robot manipulators

   77 citations · 2005
7. 7
   Efficient grounding of abstract spatial concepts for natural language interaction with robot platforms

   59 citations · 2018
8. 8
   Introspective classification for robot perception

   42 citations · 2015
9. 9
   Living with the singularity of robot wrists

   42 citations · 2005
10. 10
    Multimodal estimation and communication of latent semantic knowledge for robust execution of robot instructions

    40 citations · 2020

Key Collaborators

Contact & Links