Max Robotics — Global Service Robot Directory + US Certification

About

Dongdong Zhao is a rising leader in data-driven control and modeling for nonlinear robotic systems. His research centers on developing advanced computational frameworks that bridge Koopman operator theory, deep learning, and model predictive control to enable autonomous systems to learn and adapt without prior knowledge of their dynamics. Zhao’s major contributions include the Kalman–Koopman LQR (KKLQR) approach, which integrates optimal Koopman eigenfunctions with neural networks for robotic control, and the Deep Bilinear Koopman MPC (DBKMPC) method, which achieves both the computational speed of linear models and the predictive accuracy of nonlinear ones—each garnering 25 citations. His most recent work introduces the Kolmogorov–Arnold Transformer Koopman model (KATKM), a novel architecture that combines Kolmogorov–Arnold networks with transformers to capture complex spatiotemporal dynamics. Zhao also proposed a data-driven linear parameter-varying MPC (DDLPVMPC) that autonomously constructs system models via sparse regression. His work is highly impactful for students and researchers in robotics, control theory, and machine learning, offering practical, scalable solutions for real-time control of unknown nonlinear systems.