Zlatko Kolev

Papers

Total Citations

H-Index

About

Zlatko Kolev is a robotics researcher whose work lies at the intersection of learning from demonstration, motion planning, and human-robot interaction. His primary research focuses on developing algorithms that enable robots to learn robust trajectory distributions from human demonstrations, particularly for challenging applications like assisted teleoperation. Kolev’s key contribution is addressing a critical limitation in traditional learning from demonstration approaches: the inability to handle suboptimal demonstrations or adapt to drastic environmental changes after the initial training phase. His most cited work, “Learning Trajectory Distributions for Assisted Teleoperation and Path Planning” (2019, 19 citations), proposes a framework that learns probabilistic trajectory models capable of generalizing beyond imperfect human inputs. This is further extended in his second highly cited paper, “Reinforcement Learning of Trajectory Distributions” (2019, 11 citations), which combines reinforcement learning with trajectory distribution models to improve performance in real-time teleoperation tasks. By tackling the practical challenges of real-world human-robot collaboration—where user inputs are often noisy or suboptimal—Kolev’s research has direct implications for safer and more intuitive assistive robotics, making his work valuable for researchers in human-robot interaction and autonomous manipulation.