Special Sessions

---

Newest Advances in Systems and Control from Recent DCSD CAREER Awardees

Special Session TuAMT4
Date & Time: October 3rd Tuesday, 10:00-11:30AM
Location: Sapphire

Special Session WeAMT6
Date & Time: October 4th Wednesday, 10:00-11:30AM
Location: Balcony Room

Session Organizers

Robert G. Landers, University of Notre Dame
Jordan Berg, National Science Foundation

Abstract
This special session brings together recent CAREER awardees from the Dynamics, Control and Systems Diagnostics program at the National Science Foundation to discuss their research. You will hear about the latest advances in a wide variety of topics covering both fundamental understanding and cutting-edge applications. The format of this session will provide the audience ample time for questions and for the entire group to explore future directions in the field of systems and controls.

Session Talks

"Geometric Understanding of Locomotion"

Ross Hatton, Oregon State University

Abstract
Locomotion is the process by which an animal or robot uses its interactions with the surrounding environment to convert internal shape changes into net motion through the world. The locomotion process is in general highly nonlinear; in many cases, however it is fundamentally geometric: the amount of displacement generated over each gait cycle determined by the area enclosed by the gait trajectory, and the cost of completing this motion is determined by the length and curvature of the trajectory. In this work, we develop tools based on these geometric principles that provide insight into the optimal motions available to systems with different body plans and in different physical regimes, allow for fair comparison between disparate locomoting systems, and support data-driven modeling and analysis of gaits.

"The Effects of Viscoelasticity and Heterogeneity in Shaping Dynamic Response"

James Gilbert, Purdue University

Abstract
This presentation details recent research utilizing both material viscoelasticity and heterogeneity to tailor dynamic response using viscoelasticity. The viscoelasticity is shown through the response of a lumped parameter system that consists of a lumped parameter oscillator attached to a viscoelastic element. The parameters of the systems can be manipulated to represent a system with viscoelastic elements that has various static bifurcations. In terms of steady state response due to external excitation, particular attention is placed on the system parameters that yield a bistable system with a saddle node bifurcation. Here we show that the presence of viscoelastic effects exhibits a deviating response to a system when compared to an equivalent viscous system and effects the results to chaos. The latter is shown through a Melnikov analysis. The second part of the talk focuses on the heterogeneous design of mechanical metamaterials that enable stiffness programmability. The prototype is composed of an elastomer matrix containing diamond-shaped cavities that are selectively confined along their diagonals by semi-rigid plastic beam inserts. The system can admit 2¹² possible force-displacement relations; however, by admitting only symmetric patterns, the number of relationships reduces to 72. The talk will conclude with thoughts on how to combine the results to create systems that can exploit both phenomena to create a desired dynamic response.

"Formalizing Stability in Learning-based Policies with Switched Gaits for Legged Locomotion"

Koushil Sreenath, University of California, Berkeley

Abstract
Learning-based approaches such as reinforcement learning have demonstrated remarkable robustness and adaptability to model uncertainty in achieving legged locomotion. However, a known limitation of learning-based policies is that they lack formal guarantees. In this presentation, I will demonstrate that simple techniques from nonlinear control theory can be employed to establish formal stability guarantees for reinforcement learning policies. We will also see what stability means for motions with switched gaits.

"Fast and Accurate Battery Biometrics: How to Get Them through Optimal Excitation Design?"

Xinfan Lin, University of California, Davis

Abstract
Testing/diagnostics of lithium ion battery to determine its health and remaining useful life is critical for informing actions, such as maintenance, recycling, and repurposing, to extend the battery system life cycle and enable a circular battery economy. Current testing practice mostly relies on data from simple empirical input excitations, such as constant charge/discharge and pulse current, which takes significant amount of time and drives up the cost of testing. We are interested in input excitation design to achieve the optimal balance of testing time and accuracy by combining control, estimation, and machine learning. We first investigate the design criteria that are directly related to the accuracy of estimating battery health and are feasible for the optimization of long input sequence. We then explore both model-based and reinforcement learning approaches to solve the formulated input optimization problem. We will demonstrate a reinforcement-learning trained input control policy that could achieve fast, robust, and accurate testing, when informed by real-time state estimation.

"Modal Analysis of the Human Brain Dynamics"

Mehmet Kurt, University of Washington, Seattle

Abstract
Traumatic brain injury (TBI) is a major cause of death in the U.S., with at least 2.8 million people diagnosed annually and about twice as many unreported cases. Understanding human brain mechanics associated to TBI represents a crucial step towards improving protective devices and patient outcomes. For this reason, progress in medical imaging acquisition and processing has accompanied the development of high-fidelity structural models of the brain, using finite element or material point methods. A crucial point to consider is that the brain is a complex biomechanical system with an intricate geometry, nonuniform interfacial boundary conditions, and significantly inhomogeneous and nonlinear material properties. Most recent structural brain models feature strong nonlinearity sources, including large displacement and hyperelastic properties of brain tissues. Peculiar nonlinear dynamical behaviors have been evidenced in several studies of head impacts using these models. Prior studies analyzed the sensitivity of the human brain to deformation and found localized modes and multimodal behavior in deep regions of this tissue and observed shear-wave redirection near the falx cerebri and the tentorium cerebelli. Due to the complex design of the models and to their significant computation time, however, the exact underlying nonlinear mechanisms could not be unraveled yet. In this context, the objective of the present study is to investigate the fundamental nonlinear dynamics of the brain by using nonlinear modal analysis techniques. In this talk, I will provide insights about the spatiotemporal dynamics of the human brain by using a combination of finite-element modeling and medical imaging. I will also introduce the concept of “amplitude-resonance”, a fundamental novel concept to study the dynamics of biological material, including the human brain.

"Balancing Human Needs and Environmental Conservation in Water Systems: a Boundary Control Approach"

Mamadou Diagne, University of California, San Diego

Abstract
Efficient water management strategies are of paramount importance in safeguarding our planet from potential climate catastrophes. To avert potential climate disasters, protect the vital relationships between water, food, energy, population growth, and maintain a delicate balance of living ecosystems, we must promote and implement efficient water management strategies at local, national, and global levels. Often, the extraction of water for various purposes, whether through gates or dams manipulation, can significantly impact the stability and health of ecosystems in the surrounding environment. The alteration of the flow regimes does not only change the natural flow in rivers but can cause substantial losses of natural habitats, degrade water quality and temperature while inducing huge disruptions in diverse ecological processes that sustain biodiversity. The focus of this talk is on addressing the challenges associated with managing water resources in human-engineered river courses. The central problem under consideration involves spatiotemporal dynamics describing the evolution of both water and sediment layers in a river powered by gate actuation mechanisms. I will present results that\ show the exponential stabilization of rivers’ and reservoirs’ sedimentation process by defining a boundary control problem governed by the Saint-Venant Exner model. PDE backstepping boundary control is employed to effectively design gate operating conditions that ensure exponential stabilization of supercritical flow regime. Moreover, I will illustrate how event-based control maintains its fundamental nature as an actuation resource-aware control approach, especially when addressing gate control challenges.

"Enabling Human-Aware and Responsive Automation through Cognitive State Modeling and Estimation"

Neera Jain, Purdue University

Abstract
Across many sectors, ranging from manufacturing to healthcare to military operations, there is growing interest in the potential impact of automation, and more broadly AI, that is truly collaborative with humans. Realizing this impact, though, rests on first addressing the fundamental challenge of designing automation to be aware of, and responsive to, the human with whom it is interacting. In this talk, I will motivate and describe the goals of the research outlined in my NSF CAREER award, namely to establish a cohesive and generalizable framework for control-oriented modeling and online estimation of the human’s cognitive state. I will especially focus on the design and implementation of the novel human subject study we have designed to collect the necessary human data for this research.

NSF Funding Information Session

Special Session TuPMT4
Date & Time: October 3rd Tuesday, 1:30-3:00PM
Location: Sapphire

Session Organizers

Alexander Leonessa, National Science Foundation
Jordan Berg, National Science Foundation

Abstract
Dr. Alexander Leonessa and Dr. Jordan Berg will be presenting information about the relevant funding programs at NSF, followed by Q&A and discussions with the attendees.

Rapid Fire Sessions

Regular Session TuRFAMT5
Date & Time: October 3rd Tuesday, 10:00-11:30AM
Location: Turquoise

Special Session TuRFPMT5
Date & Time: October 3rd Tuesday, 1:30-3:00PM
Location: Turquoise

Description
The session would begin with back-to-back oral presentations of duration 5 minutes each. Once oral presentations are over, a poster session will immediately follow in the same session room.

Poster Sessions

Poster Session TuCP
Date & Time: October 3rd Tuesday, 9:30-10:00AM
Location: South Concourse

Poster Session WeCP
Date & Time: October 4th Wednesday, 9:30-10:00 AM,
Location: South Concourse

Description
Held during the morning coffee break, this session will allow contributors to present late-breaking research results in modeling, estimation, and controls. The session on Tuesday will focus on Manufacturing Controls, while the session Wednesday will cover other areas of modeling, estimation, and controls.

Rising Stars Sessions

Networking Session, Wednesday October 4th, WeAMT4, 10:00-11:30 AM, Sapphire
Special Session, Wednesday October 4th, WeRSPMT4, 2:00–3:30 PM, Sapphire
Special Session, Wednesday October 4th, WeRSEVT4, 4:30-6:00 PM, Sapphire
Special Session, Thursday October 5th, ThRSAMT4, 10:00–11:30 AM, Sapphire
Special Session, Thursday October 5th, ThRSPMT4 , 1:00-2:30 PM, Sapphire

Session Organizers

Satadru Dey, The Pennsylvania State University
Hossein Rastgoftar, University of Arizona

Description
The ASME Dynamic Systems and Controls Division (DSCD) Rising Stars special session will provide an opportunity for the young members of the community to present their research work/portfolio to a larger audience. This session is designed to help kick-start the careers of postdocs and senior Ph.D. students nearing graduation. The selected individuals will give 13 -minute-long presentations (with 5-minute Q&A) during special DSCD Rising Stars special sessions to discuss their recent and most high impact research.

Careers after PhD: Possibilities and DEI Perspective

Special Session ThAMT6
Date & Time: October 5th Thursday, 10:00-11:30 AM
Location: Balcony Room

Session Organizers

Satadru Dey, The Pennsylvania State University
Hossein Rastgoftar, University of Arizona

Introduction
This session is aimed at young members of the controls community (for example, senior graduating students and postdocs) to make them aware of various future career options. The session will focus on various career options and discuss sets of challenges associated with them. On the other hand, representation of minority groups remains low in engineering disciplines including controls. It is evident that the Under- Represented Minority (URM) groups face unique challenges in engineering careers. This special session will address these two aspects and provide an opportunity for the young members of the community to engage in conversations with representatives from various career paths – including academia, industry, and national labs.

Organization
This event will have two parts: Total 90 minutes (45 minutes for each part) Part 1 will discuss various possibilities and career opportunities in academia, industry, and national labs. This session will be a 45-minute session where panel members will share their experiences, followed by answering specific questions from the audience. Part 2 will start right after Part 1 ends. This will also be a 45-minute session. The same panelists will now focus on specific challenges faced by URM communities in engineering careers. They will first share their experiences, followed by answering specific questions from the audience

Panelists

• Academia: Prof. Juan Ren, Associate Professor, Iowa State University
• Industry: Dr. Nikolai Smolyanskiy, Director of Deep Learning for NVIDIA's Automotive group

Workshop for Middle and High School Students: Learning about Feedback Principles

Sponsor

AACC Technical Committee on Control Education, in Cooperation with the IFAC and IEEE CSS TCs on Control Education

Date & Time: October 5th Thursday, 10:00-1:00 PM,
Location: Turquoise (Lunch will be provided)

Organizer

Dr. Daniel Abramovitch, Agilent Technologies, A2C2 EdCom

Introduction
Self-driving cars, delivery drones, temperature control, wind turbines, bioreactors for medicine production, artificial limbs, and robots used in manufacturing and cleaning floors all rely on one hidden technology: human-built feedback.

For a brief video intro, see: here

Feedback as a phenomenon is everywhere: in our actions, in our bodies, and in the natural world around us. The cycle of “measure, compute, decide, adjust, and repeat” is fundamental to everything we do and everything we learn. Humans use their senses and brains to adjust their actions based on feedback. When people help other people learn to better use feedback, it is called coaching. When humans try to teach machines how to use feedback in an autonomous way, that’s where we come in. So, how do we do it? Well, these days, the way we teach a machine how to move safely around the world and interact with humans, to cook rice automatically, or control the room temperature is through software. In many ways this software is far simpler than a search engine or video game, but it has some extra constraints and stringent requirements made necessary by interacting with and adjusting the real world. Our software has to get good measurements from sensors, compare those to our desired readings, decide what to do about any errors, and then take action to adjust things for the better. Oh, and these actions have to be in the “Goldilocks” zone of not too much and not too little, and they have to be done on time (not too late). In other words, we need to apply the right feedback principles to guide us in writing the proper code.

Preliminary Schedule

• 10:00 AM: Welcome and Schedule (Daniel Abramovitch)
• 10:10 AM: Drone's to the Rescue! (Kam Leang)
• 10:30 AM: What’s a control system and why do I care? (Daniel Abramovitch)
• 11:30 AM: Bio Break, and Grab Lunch
• 12:30 PM: Hands Up Activity
• 1:00 PM: Finish Up

ASME Dynamic Systems and Control Division: Past, Present, and Future

Special Session, WeEVT2
Date: Wednesday, October 4th
Time: 4:00-5:30 PM
Location: Cobalt in Tahoe Blue Center
Also Online : Zoom Link

Session Organizers

Jingang Yi, Rutgers University
May-Win Thein, University of New Hampshire

Description
This year marks the 80th anniversary for the ASME Dynamic Systems and Control Division (DSCD). The purpose of this special session is to bring together a group of Division researchers in academia, industry, and government agencies and labs to trace the 80-year history of the DSCD, highlight significant events, show how the Division has evolved and grown over the years, and also provide insight and perspectives into the future technology developments and Division directions in system dynamics and control.

Moderator

Speaker/Panelists