Swagatika Panda

International Institute of Information Technology, Hyderabad

Papers

Total Citations

H-Index

About

Swagatika Panda is a roboticist whose research focuses on enabling robots to manipulate objects in cluttered environments by understanding the spatial semantics of their surroundings. Her key contributions lie in learning and predicting "support orders"—the hierarchical relationships between objects that determine which items must be moved first to safely access a target without causing a cascade of falls or damage. In her seminal 2013 paper, "Learning support order for manipulation in clutter" (24 citations), Panda introduced a framework for robots to infer these support relationships from single-view observations, a critical step for autonomous grasping in messy real-world settings. She extended this work in her 2016 study, "Single and Multiple View Support Order Prediction in Clutter for Manipulation" (23 citations), demonstrating how integrating multiple viewpoints improves prediction accuracy and robustness. By bridging computer vision and manipulation planning, Panda’s research has laid foundational groundwork for safer, more intelligent robotic interaction in unstructured environments—impacting fields from warehouse automation to assistive robotics. Her work continues to inspire new approaches in semantic scene understanding and dexterous manipulation.