Krishan Rana
Papers
9
Total Citations
101
H-Index
6
About
Krishan Rana is a robotics researcher whose work sits at the intersection of deep reinforcement learning, classical control theory, and robot navigation. His research is primarily focused on developing hybrid control strategies that combine the reliability of traditional controllers with the adaptability of learning-based methods — a challenge central to deploying robots in real-world, unstructured environments. Rana's most influential contribution, Bayesian Controller Fusion (BCF), elegantly merges hand-crafted control priors with model-free reinforcement learning, improving sample efficiency and enabling safer sim-to-real transfer — a persistent bottleneck in robotics research. This line of work, which forms the foundation of his doctoral thesis, has accumulated over 28 citations across multiple publications. His Multiplicative Controller Fusion and Residual Reactive Navigation frameworks further demonstrate his commitment to practical, deployable robot learning. More recently, Rana has expanded into language-grounded robotics and visual navigation. SayPlan (2023, 30 citations) showcases his ability to leverage large language models alongside 3D scene graphs for scalable task planning across complex multi-room environments, while RoboHop introduces a novel segment-based topological mapping approach for open-world navigation. Together, his body of work reflects a researcher steadily bridging the gap between theoretical machine learning and physically grounded robotic systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4Leveraging control priors in deep reinforcement learning for robotics7 citations · 2023
- 5
- 6
- 7
- 8
- 9