Lujie Yang

Papers

Total Citations

105

H-Index

About

Lujie Yang is a robotics and control theory researcher whose work sits at the critical intersection of deep learning and formal stability guarantees for dynamical systems. Their most influential contribution, "Lyapunov-stable neural-network control" (2021), has garnered nearly 100 citations and addresses one of the central challenges in modern robotics: while deep reinforcement learning has proven remarkably effective at synthesizing neural-network controllers, these systems have historically lacked rigorous theoretical guarantees. Yang's work directly confronts this gap by integrating Lyapunov stability theory with neural network design, providing a principled framework for certifiable control. Beyond neural approaches, Yang has also advanced the application of Sums-of-Squares (SOS) optimization for nonlinear dynamical systems, demonstrating in a 2023 study that SOS methods can synthesize dynamic controllers with provably bounded suboptimal performance for challenging platforms such as cart-poles and quadrotors. Together, these contributions reflect a coherent research vision: making autonomous control systems not only performant but mathematically verifiable. Yang's work is particularly valuable to researchers and engineers developing safety-critical robotic applications where stability guarantees are non-negotiable.