Dhaivat Bhatt

Université de Montréal, Robotics Research (United States)

Papers

Total Citations

H-Index

About

Dhaivat Bhatt is a robotics researcher whose work lies at the intersection of active perception, probabilistic navigation, and autonomous decision-making. His most influential contribution, "Deep Active Localization" (2019, 40 citations), pioneers an end-to-end differentiable framework for active localization—enabling robots to intelligently generate actions that maximally disambiguate their pose within a reference map. This work moves beyond traditional information-theoretic criteria and hand-crafted perceptual models, offering a learning-based approach that has become a reference point for researchers tackling the challenge of robot self-localization under uncertainty. Bhatt further advances the field with his work on "Probabilistic obstacle avoidance and object following: An overlap of Gaussians approach" (2019), which introduces a novel collision avoidance strategy that explicitly accounts for uncertainty in both the agent’s and obstacles’ states. By demonstrating that entropy-based measures can be repurposed for safe navigation, he provides a principled framework for robots to follow objects while avoiding collisions in dynamic environments. Bhatt’s research is particularly notable for its integration of deep learning with probabilistic reasoning, offering practical solutions for autonomous systems operating in the real world. His contributions continue to inspire students and researchers working at the frontier of active perception and safe robot navigation.