Organizational Unit:

George W. Woodruff School of Mechanical Engineering

Permanent Link

https://hdl.handle.net/1853/70728

Parent Organization

Organizational Unit

College of Engineering

Includes Organization(s)

Organizational Unit

Advanced Mechanical Bipedal Experimental Robotics Lab

Organizational Unit

Computational Reactor and Medical Physics Laboratory

Organizational Unit

Cryo Lab

Organizational Unit

Environmentally Conscious Design and Manufacturing

Organizational Unit

Fusion Research Center

Show 4 more

ArchiveSpace Name Record

https://finding-aids.library.gatech.edu/agents/corporate_entities/120

Full item page

Publication Search Results

Now showing 1 - 10 of 1807

Magnetic Steering to Save Sight: Trabecular Meshwork Cell Therapy as a Treatment for Primary Open Angle Glaucoma

(Georgia Institute of Technology, 2023-12-12) Bahrani Fard, Mohammad Reza

Glaucoma, which affects almost 80 million people worldwide, is the main cause of irreversible blindness. The most common type, primary open angle glaucoma (POAG), causes gradual loss of vision by damaging retinal ganglion cells. The major risk factor for POAG is high intraocular pressure (IOP). Current clinical treatments for POAG aim to reduce IOP, but they often have low success rates. The trabecular meshwork (TM) is a key regulator of IOP and has been shown to undergo significant changes in POAG including a loss of cells. This motivates the regeneration or restoration of the TM as a potential treatment for POAG. While TM cell therapy has shown promise in reversal of POAG pathology, previously-developed cell delivery techniques have resulted in poor cell delivery efficiency which elevates the risk of tumorigenicity and immunogenicity and undermines therapeutic potential. In addition, a lack of comprehensive characterization of the treatment effects in an appropriate POAG model is a roadblock to clinical translation. We here tackled these shortcomings by: 1) using an optimized magnetic delivery method to significantly improve the specificity and efficiency of delivery of cells to the mouse TM, in turn reducing the risk of unwanted side-effects, and 2) employing this optimized method to test the therapeutic capabilities of two types of cells in a mutant myocilin mouse model of ocular hypertension, characterizing the morphological and functional benefits of the treatment. The central hypothesis of this work is that an optimized magnetically-driven TM cell therapy can lead to long-term clinically significant levels of IOP reduction while minimizing the risks associated with unwanted off-target cell-delivery. This work resulted in the development of a novel magnetic TM cell therapy technique which outperformed those used previously. Employing this technique proved adipose-derived mesenchymal stem cells (hAMSC) and induced pluripotent stem cells differentiated towards a TM phenotype (iPSC-TM) to be effective in IOP lowering. Mesenchymal stem cells showed superior efficacy by stably lowering the IOP by 27% for 9 months, accompanied by increased cellularity in the conventional outflow pathway. These findings, bring magnetic TM cell therapy one step closer to clinical translation.
The Interaction of Passive and Active Ankle Exoskeletons with Age-Related Physiological Changes to Improve Metabolic Cost

(Georgia Institute of Technology, 2023-12-08) Trejo, Lindsey H.

Difficulties with mobility were the most reported disability for those age 65 and over. It is well known that older adults are slower and less economical while walking compared to young. This is thought to be brought on by reduced ankle push off power and a redistribution of positive power generation to more proximal joints (e.g., hip). We believe muscle and tendon level changes create a structural bottleneck leading to these functional changes. Ankle exoskeletons have been shown to increase ankle push off, modify muscle-tendon dynamics, and reduce metabolic cost in young adults for a near immediate improvement in walking performance. There is a critical gap in understanding whether beneficial exoskeleton assistance strategies for younger adults will also benefit older adults and if so, what the underlying mechanism is that enables exoskeletons to reduce metabolic cost across age. My aims yielded a greater understanding of how people interact with ankle exoskeletons to modify metabolic cost. In Chapter 2, we characterize the muscle and tendon properties of young and older adults. This work is in prep to be submitted to the Journal of Biomechanics. In Chapter 3, we demonstrate how young and older adults respond to spring and motorlike exoskeletons. This work is in prep to be submitted to the Institute of Electrical and Electronics Engineers Transaction on Neural Systems and Rehabilitation Engineering. In Chapter 4, we demonstrate how spring and motorlike exoskeletons interact with the underlying calf muscle and tendon properties. These outcomes can improve the design and control of ankle exoskeletons to improve the cost of walking across age, leading to greater mobility and improved quality of life.
Learning 3D Robotics Perception using Inductive Priors

(Georgia Institute of Technology, 2023-12-07) Irshad, Muhammad Zubair

Recent advances in deep learning have led to a data-centric intelligence in the last decade, i.e. artificially intelligent models unlocking the potential to ingest a large amount of data and be really good at performing digital tasks such as text-to-image generation, machine-human conversation, and image recognition. This thesis covers the topic of learning with structured inductive bias and priors to design approaches and algorithms unlocking the potential of principle-centric intelligence for the real-world. Prior knowledge (priors for short), often available in terms of past experience as well as assumptions of how the world works, helps the autonomous agent generalize better and adapt their behavior based on past experience. In this thesis, I demonstrate the use of prior knowledge in three different robotics perception problems. 1. object-centric 3D reconstruction, 2. vision and language for decision-making, and 3. 3D scene understanding. To solve these challenging problems, I propose various sources of prior knowledge including 1. geometry and appearance priors from synthetic data, 2. modularity and semantic map priors and 3. semantic, structural, and contextual priors. I study these priors for solving robotics 3D perception tasks and propose ways to efficiently encode them in deep learning models. Some priors are used to warm-start the network for transfer learning, others are used as hard constraints to restrict the action space of robotics agents. While classical techniques are brittle and fail to generalize to unseen scenarios and data-centric approaches require a large amount of labeled data, this thesis aims to build intelligent agents which require very-less real-world data or data acquired only from simulation to generalize to highly dynamic and cluttered environments in novel simulations (i.e. sim2sim) or real-world unseen environments (i.e. sim2real) for a holistic scene understanding of the 3D world.
Investigation Of Critical Attributes For Transparency And Operator Performance In Human Autonomy Teaming (TOPHAT) For Intelligent Mission Planning

(Georgia Institute of Technology, 2023-12-07) Srivastava, Divya

AI-advised Decision Making is a form of human-autonomy teaming in which an AI recommender system suggests a solution to a human operator, who is responsible for the final decision. This dissertation work seeks to empower and most effectively utilize these human decision makers by supporting their cognitive process of judgement. We propose doing so by providing the decision maker with relevant information that the AI uses to generate possible courses of action, as an alternative solution to explaining or interpreting complex AI models. Findings indicate that this technique of supporting the human’s judgement process is effective in (1) boosting the human decision maker’s situation awareness and task performance, (2) calibrating their trust in AI teammates, and (3) reducing overreliance on an AI partner. Additionally, participants were able to determine the AI’s error boundary, which enabled them to know when and when not to rely on the AI’s advice. These and associated findings are then summarized as design guidance for increasing non-algorithmic transparency in human-autonomy teams, so that the guidance can be applied to other domains.
Automated Exploration of High-Mix, Low Volume Direct Write Design Spaces Through Artificial Intelligence

(Georgia Institute of Technology, 2023-12-06) Johnson, Marshall

Additive manufacturing has experienced considerable growth over the recent decades from technological infancy to functional use across a range of industries and applications. This growth has come with increased challenges as optimal function requires the marriage of both the proper material and geometry. This challenging optimization problem is currently solved through the intervention of AM experts to ensure proper process parameters are selected. However, the need for such interventions obstructs AM from reaching its full potential to deliver customized solutions in a high-mix, low-volume (HMLV) circumstance. Thus, there is a crucial need for solutions that accommodate frequent changeovers that are innate to HMLV without sacrificing the ability to analyze the complex properties of interest. Artificial Intelligence (AI) and its specific success with image analysis represents an opportunity assist in AM for HMLV manufacturing and ease the burden from AM experts. The proposed research will utilize direct ink write (DIW) to demonstrate the capabilities of AI to impact HMLV manufacturing in the following case studies: The first study will develop a generalizable image analysis and AI tool to classify DIW printed linear self-supporting filaments and demonstrate the tool’s multiple value to automated experimentation. The second study will apply further AI techniques to build upon the image analysis and AI tool using data-driven exploration to discover sufficient parameters to DIW print complex non-linear self-supporting geometries. The third study will utilize similar image analysis and AI methodology on the DIW printing of quantum dot (QD) ink drops, a vastly different AM application, to classify the spatial ink distribution of the printed drops. The tools constructed will lay the groundwork for AI-driven HMLV analysis on the fundamental print element scale which can be extended in future work to full functional parts.
Mesoscale Computational Analysis on the Reactivity of Heterogeneous Energetic Materials with Electromechanical Properties

(Georgia Institute of Technology, 2023-12-06) Shin, Ju Hwan

Extensive research was performed in the past to investigate how energetic materials (EM) ignite via thermomechanical dissipation produced by viscoplastic deformation, friction, and fracture. Yet, little is known regarding the other types of excitation leading to ignition. Electromechanically induced dissipation is one notable example of such mechanisms that have become increasingly vital to understand in the recent development of a novel, multifunctional stimulus capable of sensitizing and triggering ignition. To quantify and better understand how certain electromechanical properties of EMs can alter their ignition behavior, a two-step, multiphysics framework spanning multiple timescales is developed. The numerical simulations first track the development of electric field (E-field) in the material under external mechanical load over the microsecond timescale. The model uses a coupled mechanical-electrostatic framework for computing the stress, strain gradient, and E-field distributions of P(VDF-TrFE)/nAl (i.e., a composite film consisting of nanoaluminum particles embedded within PVDF-TrFE binder) possessing flexoelectric and piezoelectric properties. The attainment of sufficient E-field intensity within the material is then used as part of the input for the subsequent analysis, wherein dielectric breakdown and exothermic reaction processes are simultaneously resolved over the nanosecond timescale based on an electrodynamic-chemical-thermal framework. Here, the breakdown process is explicitly modeled as the irreversible transition of the material from its initially dielectric phase into conductive phase wherever the local E-field exceeds the dielectric breakdown strength. The transient current flow leads to resistive dissipation and temperature rise in these highly localized regions of breakdown, resulting in the formation of thermal hotspots that eventually serve as critical sites for the activation and the progression of exothermic reactions. The chemical reaction is modeled as a single-stage, forward kinetic process involving the catalyzed decomposition and direct pyrolysis of the PVDF-TrFE binder, followed by the exothermic fluorination of the Al particles. The reaction rates are characterized using a temperature-dependent Arrhenius equation. The species transport (i.e., Al, PVDF-TrFE, and products) is modeled as diffusion and advection driven by the pressure gradient. The analyses focus on the effects of load intensity and microstructural attributes, such as Al particle size, particle volume fraction, void size, and porosity level, on the mechanical-electrical-chemical-thermal processes with the determination of conditions for ignition being of particular interest. Further, the individual contributions of the electromechanical properties to the overall ignition behavior are delineated using poled (flexoelectric and piezoelectric) and unpoled (flexoelectric only) specimens. The ignition times (i.e., minimum time required for ignition to occur) of poled specimens are found to be ~10% shorter than those of unpoled specimens, consistent with the accompanying experiment. Smaller particle and void sizes also promote a more rapid ignition. Similarly, higher particle volume fraction is shown to lead to a quicker ignition event, owing to the intensified flexoelectric behavior at the macroscopic level and higher chemical energy stored in the aluminum. The aforementioned framework enables a systematic establishment of the material behavioral trends and the microstructure-reaction relations.
A Skin-Like Sternal Patch to Monitor Autonomic Tone During Cognitive Stress and Sympathetic Arousals in Disordered Sleep

(Georgia Institute of Technology, 2023-12-05) Zavanelli, Nathan G.

The central focus of this thesis is the development of skin-like wearable electronics and sensors that seamlessly integrate with the human body and provide hospital quality physiological monitoring and diagnostics in a simple, minimally obtrusive platform. One of the most poignant tragedies in modern medicine is that many pathologies with highly effective treatments remain undiagnosed, especially in marginalized communities. This suffering is fueled by a systemic failure in current diagnostics techniques: one the one hand, hospital grade in lab tests are expensive, low throughput, and ill-suited for continuous monitoring; on the other, wearable electronics are fundamentally limited by rigid mechanics and wired interfaces that prevent conformal skin contact, leading to poor signal quality and degraded long-term wearability. To address this critical shortcoming, I have conducted analytical, computational, empirical, and human subjects studies in soft materials and interfaces to enable a new class of wearable, wireless devices and sensors with mechanics finely tuned to transduce electrical, mechanical, and optical bio-signals from the human body. Whereas most medical research benefits only the most advantaged, my work is targeted specifically to the unique needs of marginalized communities, providing advanced diagnostic solutions to tackle some of the most pressing medical diagnostics challenges, both here in the United States and around the world. Specifically, this thesis introduces a wireless, fully integrated, soft patch with skin-like mechanics optimized through analytical and computational studies to capture seismocardiograms, electrocardiograms, and photoplethysmograms from the sternum, allowing clinicians to investigate the cardiovascular response to acute changes in autonomic state. This system demonstrated exceptional promise in detecting sleep apneas, classifying sleep stage, identifying system vasoconstriction, and quantifying cognitive stress in trials with human subjects. This thesis focuses on the fundamental studies in soft materials, flexible electronics, signal processing and machine learning that enables these potentially transformative healthcare outcomes.
Generative Design Using Deep Learning Methods for Functionality and Manufacturability

(Georgia Institute of Technology, 2023-11-28) Wang, Zhichao

Digital manufacturing aims to optimize the entire design-to-manufacturing workflow, seamlessly integrating cutting-edge technologies to boost efficiency, precision, and agility. To enable digital manufacturing, one must achieve the seamless implementation and integration of multiple critical functions, and two of them, i.e., manufacturing process selection and generative design are the focus of this research. To aid the initial design process, automated manufacturing process selection capabilities should be developed that make recommendations early before all part shape details including quality specifications are added. The manufacturing process selection procedure will always consider shape information and may or may not include material information and other manufacturing information such as part quality requirements. Selection of manufacturing processes can be achieved through either direct classification of processes based on part shape and other requirements, or by searching for similar designs and extracting manufacturing insights from previously manufactured similar designs. We have trained neural networks on different representations of 3D designs, i.e., triangle mesh, point cloud, multi-view images and voxel, to extract shape features. These extracted shape features could be used directly for manufacturing process classification, or they could be combined with material information to select similar designs and determine manufacturing processes. Our results show that these trained models can accurately predict manufacturing processes based solely on shape information, or they can select suitable manufacturing processes through similarity search using shape and material information. During the process of adding part shape details, two design procedures: design for functionality (DFF) and design for manufacturing (DFM) are relevant. During DFF, the objective is to improve the functionality of the designed part. In DFM, functional design is utilized as input to improve its manufacturability. In the end, DFF and DFM can form a loop to continuously improve design functionality and manufacturability till the solution converges. Conditional generative adversarial networks (cGANs) were trained in these two modules, DFF and DFM, to improve the functionality and manufacturability of designs. DFF directly generated functional designs from random noise and the trained cGAN could be extended to different load and boundary conditions, while DFM modified existing unmanufacturable designs to generate their corresponding counterparts. The two modules were combined in an iterative loop, i.e., the output of DFF would be the input of DFM and the output of DFM would utilized to train DFF, to directly generate designs with both good functionality and manufacturability. This research is proposed to help automate the design and manufacturing processes in digital manufacturing so that an automatic generative design method for manufacturing can be realized.
Multimodal Vibration Damping of Circular Symmetric Structures Coupled to an Analogous Electrical Network

(Georgia Institute of Technology, 2023-11-14) Luo, Alan

Multimodal vibration attenuation can be achieved through the coupling of an analogous passive electrical network that share the approximate modal properties of the mechanical structure. By tuning the electrical network parameters to produce the same modal properties as the governing equations of the mechanical system, the coupled system can achieve a significant vibration attenuation because the modes of the network and the structure match in both frequency and spatial domain. Previous research has demonstrated the feasibility of this concept in rods, bars, beams, plates, and curved beam structures, using piezoelectric patches to couple the structure to an electric network with identical plate-like and beam-like mode shapes. This manuscript revisits previously developed analytical models and experiments to improve them while extending the analogous network concept from plates and beams to circular symmetric shells such as rings. When revisiting the analogues for beams and plates, improvements will be made to the networks to account for shear stiffness and rotary inertia, which will more accurately model thicker cross section structures. Next, an analytical model will be developed that defines the in-plane dynamics of a thin circular ring and its analogous electrical network. This will be done first using the classical extensional equations of a thin circular ring, which will then be simplified using the inextensional equations. The analytical solutions to the in-plane multimodal vibration of rings will be verified through experimental and numerical methods. Further studies will combine both the shear stiffness and rotary inertia network enhancement with the in-plane and out-of-plane analogous network to develop a comprehensive method for attenuating all types of vibration modes exhibited by rings. Finally, the feasibility of extending the analogous concept to thin circular cylindrical shells will also be explored by combining known analogous networks.
Interpretable Artificial Intelligence for Personalized Human-Robot Collaboration

(Georgia Institute of Technology, 2023-08-30) Paleja, Rohan

Collaborative robots (i.e., "cobots") and machine learning-based virtual agents are increasingly entering the human workspace with the aim of increasing productivity, enhancing safety, and improving the quality of our lives. These agents will dynamically interact with a wide variety of people in dynamic and novel contexts, increasing the prevalence of human-machine teams in healthcare, manufacturing, and search-and-rescue. Within these domains, it is critical that collaborators have aligned objectives and maintain awareness over other agents' behaviors to avoid potential accidents. In my thesis, I present several contributions that push the frontier of real-world robotics systems toward those that understand human behavior, maintain interpretability, communicate efficiently, and coordinate with high performance. Specifically, I first study the nature of collaboration in simulated, large-scale multi-agent systems, exploring techniques that utilize context-based communication among decentralized robots, and find that utilizing targeted communication and accounting for teammate heterogeneity is beneficial in generating effective coordination. Next, I transition to human-machine systems and develop a data-efficient, person-specific, and interpretable tree-based apprenticeship learning framework to enable cobots to infer and understand decision-making behavior across heterogeneous human end-users. Building on this, I extend neural tree-based architectures to support learning interpretable control policies for robots via gradient-based techniques. This not only allows end-users to inspect and understand learned behavior models, but also provides developers with the means to verify control policies for safety guarantees. Lastly, I present two works that deploy Explainable AI (xAI) techniques in human-machine collaboration, aiming to 1) characterize the utility of xAI and its benefits towards shared mental development, and 2) allow end-users to interactively modify learned policies via a graphical user interface to support team development.