Myoelectric control schemes aim to utilize the surface electromyography (EMG) signals which are the electric potentials directly measured from skeletal muscles to control wearable robots such as exoskeletons and prostheses. The main challenge of myoelectric controls is to increase and preserve the signal quality by minimizing the effect of confounding factors such as muscle fatigue or electrode shift. Current research in myoelectric control schemes are developed to work in ideal laboratory conditions, but there is a persistent need to have these control schemes be more robust and work in real-world environments. Following the manifold hypothesis, complexity in the world can be broken down from a high-dimensional space to a lower-dimensional form or representation that can explain how the higher-dimensional real world operates. From this premise, the biological actions and their relevant multimodal signals can be compressed and optimally pertinent when performed in both laboratory and non-laboratory settings once the learned representation or manifold is discovered. This thesis outlines a method that incorporates the use of a contrastive variational autoencoder with an integrated classifier on multimodal sensor data to create a compressed latent space representation that can be used in future myoelectric control schemes.

Muscle disorders and induced muscle atrophy impose critical risks to the well-being of an individual, limiting normal activities of daily living. Several resistance training methods have effectively reversed the progression of muscle atrophy. Weightlifting and hydrotherapy are the two widely practiced schemes for resistance training; however, there is the potential risk of excessive loads exerted on the muscles during weightlifting, and limited accessibility and cost are barriers to hydrotherapy. An alternative is using a resistance band. Some limitations include engaging multiple muscles/joints while only unidirectional resistance is feasible. To address these limitations, a VAriable Resistance Suit (VARS) was designed to provide speed-dependent, bi-directional, and variable resistance at a single joint. As a proof of concept, an elbow module of VARS was developed and validated experimentally through a pilot study [15]. This thesis aims to investigate the feasibility of modulating speed-dependent and adjustable resistance at the targeted joints using a VAriable Resistance Suit and investigate the efficacy of the Variable Resistance Suit to induce muscle hypertrophy. The pilot study shows the changes in flexor and extensor muscle activations in response to eight different levels of resistance modulated by VARS. Furthermore, the evaluation of training using VARS on muscle hypertrophy with a focus on the increase in muscle size and strength has been conducted through a prolonged study involving 12 participants. Some sections of this thesis are reused from our published conference paper which I co-author [15].

This thesis work aims to develop electrochemical sensors for sub-ppb level detection of inorganic and organic pollutants in drinking water with environmentally benign materials and processes. While traditional laboratory-based methods such as mass spectroscopy, and chromatography have been used to analyze the concentration of contaminants in drinking water, miniaturized electrochemical sensors offer a compelling alternative to those methods, enabling rapid on-site cost-effective detection of low concentrations of pollutants. In this research, a set of three-electrode sensors was designed and fabricated on a flexible substrate using a screen-printing technique. Additionally, an in-situ electrochlorination process was implemented to create the reference electrode. These sensors were utilized to precisely detect lead ions and perfluorooctane sulfonate (PFOS) in drinking water. The first set of sensors was fabricated to measure the concentration of lead ions, a toxic inorganic pollutant, in potable water. The novelty of the proposed research lies in using non-toxic, biodegradable sodium alginate grafted with 2- acrylamido-2-methyl propane sulfonic acid (AMPS) and conductive fillers for trace-level lead ion detection in water. The principle of square wave anodic square wave stripping voltammetry (SWASV) was used to determine the trace level lead ion concentration. Employing a similar approach with a different material, a PFOS sensor was developed. Utilizing chitosan, one of the sustainable and biodegradable biopolymers found in crustacean shells, rapid parts-per-trillion (ppt) level PFOS detection by electrochemical impedance spectroscopy (EIS) was demonstrated. The proposed sensors made low-cost electrochemical detection of contaminants such as lead ions and PFOS possible with eco-friendly materials and processes.

Seismocardiography (SCG) is the low frequency chest surface vibration generated by the mechanical activities of the heart. SCG has been found to have clinical utilities in diagnosis of different cardiac diseases. The first part of this study focused on the application of SCG signal in predicting hospital readmissions of the heart failure (HF) patients. Conventional machine learning and deep learning models have been developed using SCG signal acquired from the HF patients. Early HF readmissions was predicted with decent accuracies with these models. This may potentially help the clinicians to identify the patients who need special care and treatment and make timely targeted interventions. This will ensure better management of HF patients and reduce the mortality rate. One of the limitations of using SCG signal in clinical settings is its variability. To investigate SCG variability, an exercise protocol has been developed. SCG signal was acquired from the healthy subjects when they underwent the protocol. It was found that cardiopulmonary interactions may contribute to the variability in SCG signal. The study results help to better understand the source of variability which eventually may increase the clinical utility of SCG signal. Another limitation of SCG signal is that it is highly sensitive to the ambient and locomotion-induced noises. This can distort the SCG signal. Hence, removal of noises is a necessary step to use SCG in ambulatory assessment of HF patients. To encounter this problem, a healthy subject performed different maneuvers to induce few common types of noises in the SCG signal. Different signal processing techniques have been employed to remove the noises from the signal. A comparison among different techniques has been provided which may lead to developing an algorithm in the future that is capable of autodetecting noises and suppress them.

People experiencing mobility deficiencies in their lower limbs caused by genetics, injuries, diseases, etc. struggle with their physical and mental health. The goal of this research is to design an exoskeleton that will connect the upper limb (e.g., arm extension) to the ankle joint during walking movements. We advanced the first prototype of the Workshare Upper Lower Limb (WULL) by only targeting the ankle joint as the lower limb component. We found that this change would have the biggest impact on an individual's walking movements. The benefit of this research will be found in answering the question: will harnessing the kinetic energy from a person's upper limb (e.g., arm extension or arm flexion) to transfer into the ankle joint for gait assistance reduce the lower limb muscle activation during walking movements? A series of experiments were run to test the efficacy of the wearable device. Six participants were fitted to the device and six electromyography (EMG) sensors to track the muscle activation during a comfortable walking pace. This gait analysis study used pressure insoles to calculate ground reaction forces and multiple IMUs to track the individuals' limbs and joints kinetic motion. The overall effectiveness of the device was explored based on the data collected in this study. This device decreased muscle activation of the gastrocnemii medialis and increased the anterior deltoid activation. These results support the goal of the experiment to utilize the upper limbs (anterior deltoid) to assist the lower limbs (ankle joint) during walking.

To confirm computer simulations and Computational Fluid Dynamics (CFD) analysis, benchtop experiments are needed with a fluid that mimics blood and its viscoelastic properties. Blood is challenging to use as a working fluid in a laboratory setting because of health and safety concerns. Therefore, a blood analogue is necessary to perform benchtop experiments. Viscosity is an important property of fluids for modeling and experiments. Blood is a shear thinning fluid, so it has a decreasing viscosity with higher shear rates. This project seeks to create a blood mimicking fluid for benchtop laboratory use. Numerous fluids with different combinations of water, glycerin, and xanthan gum were created to mimic the shear thinning property of blood at different hematocrit levels. Since the amount of xanthan gum is very small, an analytical balance was used. To mix the solution, an immersion blender and a heat circulator were used. The data were obtained from 10-90 torque percent, which is the range over which the rheometer is accurate, so the exact ranges of shear rate tested depended on the test fluid. The created solutions were compared to blood at the equivalent hematocrit and previously performed tests.The three different equivalent hematocrits all produced results similar to viscosities of blood. The results were similarly representative of blood at different equivalent viscosities for the 0.0075% xanthan gum and the 0.075% xanthan gum by weight. The solutions were able to mimic the shear thinning behavior of blood at different equivalent hematocrits. The fluids with 0.075% xanthan gum and 50% water and 50% glycerin is a better representative than the fluids with 0.075% xanthan gum and 60% water and 40% glycerin.

No description available.

The goal of this thesis is to develop a soft multiple-degree-of-freedom (multi-DOF) load cell that is robust and light weight for use in robotics applications to sense three axes of force and a single axis of torque. The displacement of the magnet within the elastomer changes the magnetic flux density which is sensed by two 3-axis Hall effect sensors. Experimental measurements of magnetic flux density within the area of interest were used to formulate analytic expressions that relate magnet field strength to the position of the magnet. The displacement and orientation measurement and the material properties of the elastomer are used to calibrate and calculate the applied load. The ability to measure 3-DOF force and axial torque was evaluated with combined loading applied by a robotic arm (KUKA, LBR r820 iiwa). The decoupled results show the 4-DOF load cell was able to distinguish 3-axis force and 1-axis torque with 6.9% averaged error for normal force, 4.3% and 2.6% for shear force in the X and Y axis and 8.6% for the torque. The results show good accuracy for a soft multi-axis sensor that would be applicable in many robotic applications where high accuracy is not required.

With balance assessment being essential to the prediction of fall risk and detection and assessment of motor disorders, increasing clinical accessibility to objective balance metrics is essential. Work has been done validating center of mass (COM) acceleration metrics against well regarded force plate center of pressure (COP) position standards for mobile device data acquisition systems, but most research is restricted to devices being close to the COM or strapped in a harness configuration. Through the use of rigid body kinematics and the inverted pendulum model, this study develops novel methodology for calculating COM acceleration using mobile devices in arbitrary positions, as well as a novel approach to validation through direct comparison of COP position predictions to force plate measurements. Validation of this methodology included comparison of smartphone and force plate results for COM accelerations and COP positions. Results showed good analysis performance for both approaches during subject intentional swaying, but results were limited in cases of little motion where good balance was observed.

<p dir="ltr">The focus is on the complex nature of primary blast injury (PBI) and employs advanced simulation techniques to model the physiological impacts using a TBI-on-a-chip system. This study involves a two-way Fluid-Structure Interaction (FSI) model in ANSYS, coupling Transient Structural and Fluent modules to simulate the effects of a blast wave on brain tissue. The research explores the creation and validation of boundary conditions, such as fixed support and varying strain rates, to ensure the reliability of the experimental setup. Key findings include the non-uniform distribution of strain, which has significant implications for understanding injury mechanisms and inflammatory marker analysis. The project also provides a detailed workflow for FSI simulations, highlighting the advantages of uniform mechanical loading and its impact on experimental accuracy.</p>