Papers
15
Total Citations
438
H-Index
7
About
Rohan Chandra is a robotics researcher whose work spans autonomous navigation, deep reinforcement learning, and human-robot interaction. His research tackles some of the most pressing challenges in deploying robots in real-world environments — from unstructured outdoor terrain to crowded human-populated spaces. Chandra's most cited work, "GA-Nav" (2022, 145 citations), introduced a novel group-wise attention mechanism that enables robots to efficiently identify navigable terrain from RGB images, representing a significant advance in outdoor robot autonomy. His influential survey on deep reinforcement learning for robotics (99 citations) provides a comprehensive analysis of real-world DRL successes, serving as an essential reference for researchers bridging the gap between algorithmic promise and physical deployment. A defining theme of Chandra's research is social robot navigation — developing robots that move safely and naturally alongside humans. His work on evaluation principles for social navigation algorithms (59 citations) and hybrid geometric-learning navigation frameworks has helped standardize and advance the field. He has also contributed tools like SocialGym 2.0 for multi-robot simulation and theoretical approaches ensuring deadlock-free multi-robot coordination. With over 400 cumulative citations, Chandra's work is shaping the future of intelligent, socially aware robotic systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2Deep Reinforcement Learning for Robotics: A Survey of Real-World Successes99 citations · 2024
- 3Principles and Guidelines for Evaluating Social Robot Navigation Algorithms59 citations · 2024
- 4Deep Reinforcement Learning for Robotics: A Survey of Real-World Successes58 citations · 2025
- 5Rethinking Social Robot Navigation: Leveraging the Best of Two Worlds17 citations · 2024
- 6Principles and Guidelines for Evaluating Social Robot Navigation Algorithms15 citations · 2023
- 7
- 8
- 9
- 10