AI for Dynamical Systems and Control

3.0 credits · Graduate

This course introduces modern machine learning, neural network, and artificial intelligence methods for the modeling, prediction, estimation, and control of multivariable dynamical systems. The course emphasizes the integration of data-driven learning techniques with classical multivariable systems and control theory to address challenges arising in complex nonlinear systems for which first-principles models are incomplete, uncertain, or unavailable. Topics include supervised, unsupervised, and reinforcement learning methods; neural networks for system identification and state estimation; neural ordinary differential equations; physics-informed machine learning; optimal control; model predictive control; adaptive and learning-based control; and reinforcement learning for autonomous decision making. Applications include robotics, autonomous vehicles, aerospace systems, manufacturing, energy systems, biological systems, and cyber-physical systems.