Vishnu Dutt Sharma

Papers

Total Citations

H-Index

About

Vishnu Dutt Sharma is a robotics researcher advancing the frontier of autonomous navigation and 3D perception. His work centers on integrating predictive models into active perception systems, enabling robots to anticipate environmental uncertainty rather than simply react to it. His most-cited paper, "Pred-NBV: Prediction-Guided Next-Best-View Planning for 3D Object Reconstruction" (2023, 19 citations), introduces a framework that uses shape prediction to guide a robot’s viewpoint selection, dramatically improving reconstruction efficiency and safety. In "ProxMaP: Proximal Occupancy Map Prediction for Efficient Indoor Robot Navigation" (2023, 5 citations), Sharma tackles the challenge of navigating unknown spaces by predicting occupancy in unmapped regions, allowing robots to plan faster, less conservative paths. His most recent work, "Pre-Trained Masked Image Model for Mobile Robot Navigation" (2024, 2 citations), leverages large-scale pre-trained vision models to infer structural patterns in top-down maps, reducing reliance on exhaustive online mapping. Sharma’s contributions are notable for bridging computer vision and robotics—using learned priors to make real-time navigation more intelligent and adaptive. His research is particularly impactful for applications in search-and-rescue, warehouse automation, and autonomous exploration, where efficiency and foresight are critical.