Papers
12
Total Citations
122
H-Index
6
About
Harit Pandya is a robotics researcher whose work sits at the intersection of soft robotics, computer vision, and human-robot interaction. His most impactful contribution is a deep learning method for vision-based force prediction in soft Fin Ray grippers, using simulation data to overcome the challenges of modeling flexible, nonlinear robotic systems—a paper that has garnered 40 citations. Pandya has also made significant strides in haptic shared control, developing an intent-aware predictive guidance system that dynamically adjusts robotic assistance based on human motion predictions (18 citations). His research extends to non-prehensile manipulation, where he introduced "Push-to-See," a deep Q-learning approach that enhances instance segmentation in cluttered scenes (14 citations), and to visual servoing, where he pioneered instance-invariant frameworks for autonomous vehicle inspection using MAVs (11 citations) and addressed the challenge of servoing toward tumbling objects in space (9 citations). More recently, Pandya has explored world models for reinforcement learning with his ReCoRe framework (5 citations). His work demonstrates a consistent focus on enabling robots to perceive, manipulate, and interact with complex, dynamic environments.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5Servoing across object instances: Visual servoing for object category10 citations · 2015
- 6Image Based Visual Servoing for Tumbling Objects9 citations · 2018
- 7ReCoRe: Regularized Contrastive Representation Learning of World Model5 citations · 2024
- 8Pose induction for visual servoing to a novel object instance5 citations · 2017
- 9
- 10