Tutorial Sessions

Tutorial 1: A Tutorial on the Forgotten Signals in Control
Organizer: Daniel Y. Abramovitch [Agilent Technologies]

This tutorial session will consist of two separate tutorials that are related to each other in that the signals and methods discussed are often treated as an afterthought in feedback design, even though they often determine the limits of the control system The first tutorial will be on noise propagation through a system and analysis of that noise, with a focus on the PES Pareto method invented in the 1990s and used extensively in the storage industry since then.  Under certain assumptions, PES Pareto provides a powerful tool for isolating noise sources from loop measurements and for quantifying their effect on other signals in closed-loop.  We will discuss different methods of generating measurements to fill out the signals needed for this scheme as well as discuss how the methods can be used to predict the effects of increases or decreases in any one noise source.  Since the original tutorial, the author has shown that PES Pareto can also be used to quantify the noise covariance inputs needed for Kalman and Kalman-Bucy Filters.

Tutorial 2: Learning for Complex Dynamics and Controls
Organizers: Sidra Bhatti [Ohio State University], Qadeer Ahmed [Ohio State University]

This tutorial presents a comprehensive treatment of Learning for Complex Dynamics \& Controls, integrating machine learning (ML) with classical control theory to address the limitations of traditional model-based strategies for real-world dynamical systems. Through a cohesive presentation of foundational and advanced techniques — including deep reinforcement learning, physics-informed neural networks, transformer-based architectures, and sequence-aware models — the tutorial demonstrates how learning-based controllers can reliably optimize competing objectives such as fuel efficiency, emissions, and battery longevity without compromising safety constraints. The content is grounded in real-world case studies from ongoing research at The Ohio State University, spanning explainable energy management in hybrid powertrains, physics-informed aging- and emissions-aware control, parameter estimation in partially observable systems, and sequence-aware energy management in series hybrid trucks. The tutorial targets researchers, industry practitioners, and graduate students with a basic background in control systems or ML, equipping attendees with both the conceptual vocabulary and practical tools to advance their work at the intersection of machine learning, control engineering, safety, and efficiency.