Mirela Popa

Universität Innsbruck, Maastricht University

Papers

Total Citations

H-Index

About

Mirela Popa is a researcher at the forefront of human-robot collaboration (HRC) and robotic perception. Her work primarily focuses on enabling robots to understand and anticipate human actions, bridging the gap between reactive control and truly intuitive teamwork. A key contribution is her pioneering approach to proactive robot task sequencing, where real-time hand motion prediction allows robots to pre-plan their actions, significantly improving efficiency and safety in shared workspaces. This forward-looking method, detailed in her 2025 paper, addresses the critical need for robots that can infer human intent rather than simply react to it. In her earlier foundational work, Popa introduced "CPS: 3D Compositional Part Segmentation through Grasping," a novel framework that teaches robots to understand objects not as monolithic forms, but as compositions of semantically meaningful parts—like a handle designed for grasping. This semantic understanding is crucial for dexterous manipulation. While her citation counts are still growing, the innovative, proactive nature of her recent research positions her as a rising voice in the next generation of collaborative robotics, with clear implications for manufacturing and assistive technologies.