Ekaterina Tolstaya

California University of Pennsylvania, University of Pennsylvania

Papers

Total Citations

163

H-Index

About

Ekaterina Tolstaya is a leading researcher in multi-agent robotics and decentralized control, whose work is redefining how large swarms of robots coordinate autonomously. Her core research lies at the intersection of graph neural networks (GNNs) and multi-robot systems, where she has pioneered scalable, decentralized controllers that require only local information and communication. Her most influential contribution is the development of spatial graph neural networks for multi-robot coverage and exploration—a critical capability for inspection, search and rescue, and environmental monitoring. This work, along with her papers on learning decentralized controllers for robot swarms, has each garnered 67 citations, demonstrating significant impact in the field. Tolstaya also introduced Graph Policy Gradients (GPG), an algorithm that exploits graph symmetry among homogeneous robots to overcome the curse of dimensionality in large-scale control. Her methods bridge the gap between optimal centralized control and practical, scalable decentralized execution, offering a principled framework for controlling complex, interacting dynamical systems. Through her innovative use of GNNs, Tolstaya is enabling the next generation of autonomous swarms for applications ranging from smart grids to smart cities.