We develop unified control-communication frameworks for networked control systems. Our work is concerned with adaptive and robust designs across diverse system architectures and stability guarantees for nonlinear large-scale and infinite networks.
In a further research pillar, we advance data‑driven and learning‑enhanced control across the full modelling spectrum. For white‑box settings, we generalise the fundamental lemma to nonlinear and stochastic regimes, enabling non‑parametric identification of input–output maps with finite data. For black‑box scenarios, we embed deep neural operators within predictive‑control layers to handle high‑dimensional environments for autonomous decision making and scene understanding. Of interest are also grey‑box models that fuse first‑principles structure with learned representations, yielding interpretable and adaptive hybrid models.
In parallel, we develop modeling and optimal control frameworks using partial differential equations (PDEs), with applications spanning chemical engineering, systems biology, and cancer therapeutics. Our research integrates data-driven and DNN-based control schemes, emphasizing the synergistic interplay between deep learning and dynamical systems.
