Papers
7
Total Citations
95
H-Index
4
About
Anirvan Dutta is a pioneering roboticist whose research sits at the intersection of perception, manipulation, and autonomy. His primary focus is on developing **active visuo-tactile frameworks** that enable robots to intelligently explore and understand their environment. Dutta’s major contribution lies in creating systems that combine visual data with tactile feedback to infer critical object properties—such as shape, stiffness, mass, friction, and center of mass—in dense, unstructured clutter. His most influential work, “Active Visuo-Tactile Interactive Robotic Perception for Accurate Object Pose Estimation in Dense Clutter” (41 citations), introduces a novel *declutter graph* to model object relationships, allowing robots to accurately estimate poses even in challenging scenes. He further advanced the field with “Push to Know!” (12 citations), which employs dual differentiable filtering for parameter inference, and “Sensorless Full Body Active Compliance” (31 citations), demonstrating innovative control in parallel manipulators. Dutta’s work is not only highly cited but also practically impactful, bridging the gap between theoretical perception models and real-world robotic dexterity. His recent contributions, including predictive and cross-modal perception frameworks, continue to push the boundaries of how robots learn through interaction, making him a leading voice in autonomous robotic exploration.
Research Focus
Key Achievements
Top Papers
- 1
- 2Sensorless full body active compliance in a 6 DOF parallel manipulator31 citations · 2019
- 3
- 4
- 5
- 6
- 7