Leonidas Koutras
Papers
8
Total Citations
111
H-Index
5
About
Leonidas Koutras is a roboticist whose research sits at the intersection of learning from demonstration, non-prehensile manipulation, and compliant control. His most impactful work addresses fundamental limitations in Dynamic Movement Primitives (DMPs), a widely used framework for encoding and generalizing robot trajectories. In a pair of highly cited papers (26 and 24 citations), Koutras identified and corrected critical flaws in existing DMP formulations for both spatial scaling and orientation in Cartesian space, providing singularity-free and frame-independent solutions that have become essential references for the community. He further extended the DMP framework to handle moving goals with temporal scaling adaptation (23 citations), enabling robots to track dynamic targets without explicit motion models. Beyond trajectory learning, Koutras has made significant contributions to robotic grasping in dense clutter, developing a multi-fingered push-grasping policy (23 citations) that strategically uses non-prehensile actions to create collision-free grasp affordances. His more recent work explores passivity-based control for bimanual manipulation of large objects and dynamic obstacle avoidance for compliant robots, demonstrating a commitment to building safe, physically interactive systems. With a publication record spanning from foundational DMP theory to practical manipulation strategies, Koutras is shaping how robots learn and adapt in unstructured, human-centered environments.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Learning Push-Grasping in Dense Clutter23 citations · 2022
- 4
- 5
- 6Enforcing Constraints for Dynamic Obstacle Avoidance by Compliant Robots3 citations · 2023
- 7
- 8