Probabilistic Control

Probabilistic control aims to design controllers for systems operating under uncertainty, optimizing performance by considering the probability distributions of states and disturbances. Current research focuses on developing efficient algorithms, such as those based on expectation-maximization or iterative inference, to solve the resulting optimization problems, often leveraging techniques from uncertainty quantification and probabilistic inference. These advancements enable handling complex nonlinear systems and non-Gaussian uncertainties, improving control performance and robustness in robotics and other applications where precise modeling is challenging. The resulting methods offer a powerful framework for addressing control problems in domains with inherent stochasticity.