Anastasija Demerdjieva

Papers

Total Citations

H-Index

About

Anastasija Demerdjieva is a researcher at the intersection of robotics, control theory, and machine learning, with a primary focus on developing data-driven methods for real-time optimal control. Her most cited work, "Learning optimal controllers: a dynamical motion primitive approach" (2023), addresses the fundamental challenge of computing optimal control policies in real time—a problem that has long hindered the deployment of sophisticated robotic systems. Rather than relying on computationally expensive online optimization, Demerdjieva proposes learning (potentially sub-optimal) controllers that can be executed instantaneously, bridging the gap between theoretical optimality and practical feasibility. This work, which has garnered early citations, introduces a dynamical motion primitive framework that simplifies the learning of complex control behaviors. Her contributions are particularly valuable for applications in autonomous navigation, manipulation, and human-robot interaction, where split-second decisions are critical. By enabling robots to approximate optimal behavior without heavy computation, Demerdjieva is helping to make advanced control accessible for real-world deployment. Her research represents a promising step toward more intelligent, responsive, and efficient autonomous systems.