System Identification

System identification aims to determine a mathematical model representing a dynamical system's behavior from observed input-output data. Current research emphasizes developing robust and efficient methods for identifying both linear and nonlinear systems, employing diverse approaches such as neural networks (including Physics-Informed Neural Networks and recurrent architectures), Bayesian methods, and optimization techniques like L-BFGS-B and Expectation-Maximization. These advancements are crucial for improving control system design, enabling accurate state estimation, and facilitating data-driven modeling across various fields, including robotics, structural health monitoring, and biological systems.