Abstract

Deep unfolding is a widely acknowledged computational technique used for parameter tuning, enhancing the convergence speed of iterative algorithms. Its utility and investigation have been extensive across various domains, including wireless communications. This presentation presents a detailed overview of the speaker's research, highlighting the integration of deep unfolding in the field of control engineering. This discussion covers recent discoveries relating to the design of feedback control systems, multi-agent systems and model predictive control, with its effectiveness confirmed by simulations in the control of non-linear dynamical systems such as tank reactors, multi-link pendulums and neural networks. It emphasizes the use of deep unfolding to tackle intricate control hurdles in these various systems, enhancing control engineering progress.

Biography