62970 research outputs found
Sort by
A Novel Smart Wearable System with Edge Computing AI for Cardiac Disease Detection and Continuous Monitoring
No full textCardiac signals are critical physiological markers that provide crucial information
for diagnosing various cardiovascular diseases (CVDs). Many cardiac and cardiopulmonary diseases require monitoring through an electrocardiogram (ECG), as well as respiration, blood oxygen level, body temperature, and other physiological signals. Underlying cardiopulmonary issues can lead to serious diseases like arrhythmias, myocardial infarction, heart attack, bradypnea, tachypnea, chronic obstructive pulmonary disease (COPD), and other health concerns. Wearable devices have been useful in monitoring and tracking multi-dimensional physiological signals. Regular users do not tend to go to doctors unless they have any symptoms. Since clinical evaluation typically occurs only after symptom onset, pathologies can advance to severe stages before detection. Moreover, motion and posture variability degrade real-life monitoring continuity, and without on-device edge computing, wearables must rely on post-processing that impedes true real-time detection.
We have designed a multimodal wearable system capable of simultaneously
monitoring cardiac and respiration signals in static activities. A single printed circuit board (PCB) is developed to capture ECG and respiration. The ECG analog
front end (AFE) captures cardiac activity using dry inkjet printed (IJP) electrodes.
The AFE contains an instrumentation amplifier-based design along with
multistage filtering and amplification. An inertial measurement unit (IMU) is
used to capture the respiratory behavior in static and minor moving conditions,
such as working on an office desk. The breathing signal processing includes dynamic filtering and a robust axis selection mechanism. The firmware processes
the raw data and transmits via Bluetooth Low Energy (BLE) in real-time. The work explores the feasibility of simultaneous multimodal data acquisition in static
and indoor conditions.
We further enhanced the performance of the wearable system with multimodal
physiological monitoring of ECG, respiration, peripheral capillary oxygen saturation (SpO2), and body temperature during both static and dynamic activities and postures. The system incorporates a QRS detection algorithm and a novel single IMU based quaternion update for ECG and respiration monitoring in sitting, standing, walking, running, and cycling conditions. The firmware on the microcontroller of the wearable processes the raw data and transmits raw and
processed features via Bluetooth Low Energy (BLE) module in real-time. The
performance of the wearable system was validated with a reference commercial
device. The amplifier chip offers a high common-mode rejection ratio (CMRR) of
86 dB and a gain of 56 dB. The proposed ECG and respiration processing is validated with reference commercial device and provides 98.4% and 98.9% accuracy, respectively, for heart rate (HR) and respiration rate (RR) measurements. The work also explores the performance of the system in an ambulatory setup during indoor and outdoor conditions.
Finally, we design and integrate lightweight Artificial Intelligence (AI) models into the wearable device for ECG arrhythmia detection. Lightweight deep neural network (DNN) models, if integrated into a low-resource system on chip (SoC), can offer low-cost, low-power smart health solutions that provide more accurate
real-time analysis and reduce the computational burden on mHealth and cloud applications. The work enhances the performance of light AI models by utilizing real-time QRS detection and a custom morphological feature set for the DNN input layer. The on-chip DNN model can classify five arrhythmic beats in real-time, beat-by-beat analysis. The QRS detection block offers 99.6% accuracy for R peak detection and a low overall R peak position error (RPE) of 6.3 ms. The real-time feature extraction block offers a good correlation factor of 95.1% with reference data. This resource-constrained DNN model is 98.1% accurate in detecting arrhythmia ECG beats and takes 20 ms for inference, which makes the system reliable for real-time applications. The wearable system consumes only 5.9 mA for beat-by-beat arrhythmia detection. The research delivers a low power,
lightweight, low-latency edge-computing solution for continuous real-time
monitoring and inference across static and dynamic postures in both indoor and
outdoor settings, offering key insights into multimodal system design and data
processing for cardiopulmonary disease tracking and detection
Evaluation of Accessible Technologies to Assess Body Composition, Body Water, and Metabolism in Muscular Resistance-trained Adults
No full textA comprehensive nutritional assessment includes body composition and metabolism evaluation. This is a critical component in athletic and clinical settings. While a number of methods exist for quantifying select body composition and metabolism metrics, it is critical that valid and reliable techniques be employed. Previous works have suggested that the muscularity of individuals may influence the accuracy of body composition and metabolism estimates as compared to the general population. The purpose of the current study was to 1) determine the validity of several accessible techniques for estimating body fat percentage, fat-free mass, fat mass, total body water, and resting metabolic rate, 2) evaluate whether the fat-free mass properties of muscular adults vary from accepted cadaver reference values, and 3) assess the agreement between multiple non-invasive techniques for estimating skeletal muscle mass in muscular, resistance-trained adults.
Forty resistance-trained adults (males = 23; females = 17; mean ± SD; age = 28.4 ± 7.9 years; height = 171.9 ± 9.4 cm; body mass = 77.1 ± 12.8 kg; body fat % = 15.7 ± 5.1%; fat-free mass index = 21.9 ± 2.8 kg/m2) enrolled in the study. Participants completed two total visits to the research laboratory. The first visit was a screening session to determine each participant’s eligibility based on pre-specified resistance training history and body composition criteria. Visit two was the experimental testing visit in which multiple body composition and metabolism assessments were completed, along with a number of questionnaires.
The primary findings of the current investigation indicated that the high muscularity of the individuals in our sample likely influenced the validity of several body composition and metabolism techniques. Nonetheless, dual-energy X-ray absorptiometry and air displacement plethysmography demonstrated the best validity for estimating body fat percentage, fat-free mass, and fat mass compared to the other methods. A supine bioimpedance spectroscopy device and the Kushner bioimpedance prediction equation were also found to be valid for estimating total body water in muscular adults. While the fat-free mass fraction of total body water did not differ from reference values, other variables including density, mineral, and residual did differ, and variation was noted between males and females in our study as well. Several differences between methods for estimating skeletal muscle mass were additionally noted and require further evaluation in muscular individuals against a criterion method. Lastly, of the methods assessed for estimating resting metabolic rate, none can be recommended, as every method demonstrated unacceptable underestimation when compared to indirect calorimetry
Roadside LiDAR-based Vehicular Trajectory Repair and Lane Marking Identification for Connected and Autonomous Vehicles
No full textRoadside LiDAR has emerged as a promising sensing technology for intelligent transportation systems, offering high-resolution spatial data for tasks such as vehicle tracking, digital map generation, and traffic analysis. However, its practical deployment remains limited due to critical challenges including occlusion-induced data loss, sparse point density at long distances, and the labor-intensive nature of traditional map construction methods. Therefore, we propose a cross-modal integration framework that significantly enhances the completeness, accuracy, and scalability of roadside LiDAR applications.
First, a lightweight and low-cost lane-level mapping method is developed by aligning aerial imagery with roadside LiDAR data. Through building edge matching and a novel rotation estimation strategy, this approach enables accurate map alignment without requiring dense traffic or repeated scanning. The resulting lane-level maps demonstrate a high trajectory alignment accuracy, providing a scalable solution for large-scale digital twin construction.
Second, a trajectory supplementation method is proposed to reconstruct vehicle paths under conditions of full occlusion and limited data points. By associating fragmented vehicle IDs and identifying representative frames with the highest point density, the method restores tracking continuity even in dense urban intersections and peripheral sensing zones. This enhancement significantly improves tracking reliability and enables extended-range analysis. Lastly, the improved trajectory data is used to support shape-based vehicle classification, enabling the differentiation of vehicle types in a more accurate and consistent manner.
These contributions transform roadside LiDAR from a geometry-limited sensor into a semantically enriched, cross-modally aware platform suitable for modern ITS infrastructure. The proposed framework lays the groundwork for scalable, accurate, and low-cost deployment of LiDAR in real-world traffic environments
A Deep Dive into the Cognitive Soundscape of Flow: Finding Your Groove
No full textFlow state, characterized by optimal engagement and performance, represents a key concept in understanding human performance and cognitive resource allocation. Grounded in Csikszentmihalyi's & Sherry’s flow theory and the Limited Capacity Model of Motivated Mediated Message Processing (LC4MP), this study investigated physiological and neural correlates of flow state during a simulated driving task under different music conditions and difficulty levels. Using a 2 × 3 factorial design with 20 participants, this study examined self-selected versus non-self-selected music across three difficulty levels, testing the relationship between task switching, cognitive resource allocation, and flow state. Physiological measures included heart rate and EEG (alpha/theta power) using a 4-channel Muse 2 headband, alongside the self-report measure for flow experience. Hierarchical linear modeling revealed significant physiological changes during self-selected music: heart rate decreased (β = -5.15, p < 0.001), while alpha (β = 5829.77, p < 0.001) and theta power (β = 7637.24, p < 0.001) increased. Task difficulty also showed significant effects, with heart rate decreasing during hard (β = -6.70, p < 0.001) and moderate (β = -3.40, p = 0.001) conditions. Notably, while physiological measures showed robust changes, self-reported flow state (PO) did not reach significance. Task switching rates showed significant decreases during self-selected music (β = -0.86, p < 0.001) and hard difficulty (β = -0.61, p < 0.001), supporting the LC4MP framework's predictions about cognitive resource allocation. These findings demonstrate how task switching and cognitive resource allocation relate to flow state induction. The results highlight the importance of multiple measurement modalities and demonstrate that personal relevance (through music selection) and task difficulty significantly influence physiological and neural responses during performance. Future research should employ more comprehensive measurement approaches to better capture the complexity of flow-related neural activity and its relationship to task switching and cognitive resource allocation
Strengthening Universal Inclusion: A Teacher Training Program for Academically Advanced Students
No full textThis research investigates the efficacy of a pedagogical training program that focuses on leveraging students' talents, specifically developed for middle school instructors who educate academically advanced students. The research takes place at a junior high school in Unnamed city, which is a Title I school. The school follows an "Advanced for All" curriculum, which requires all students to take advanced material. This study focuses on the issue of instructors' lack of readiness to effectively cater to the unique learning requirements of academically gifted pupils, particularly using differentiated instruction. The researcher, who is a member of the school's teaching community, utilized insider action research techniques to perform a formative evaluation of the training program. This evaluation was carried out using semi-structured interviews. This approach facilitated an in-depth examination of instructors' viewpoints about the acquisition, integration, and implementation of pedagogical practices that focus on leveraging strengths. The study offers valuable insights into the impact of the training program on instructors' perspectives on academically gifted kids, their lesson preparation, and instructional methods. Evidence indicates that professional development programs focused on building strengths can improve instructors' capacity to stimulate and involve high-achieving students, resulting in enhanced student performance and teacher effectiveness. This study adds to the expanding collection of knowledge on strengths-based education and provides practical suggestions for implementing these training programs in comparable school settings
Profitability Impacts of Soil Health Management Practices on Cotton-Based Cropping Systems in the Texas High Plains
No full textCognizant of the serious water constraint due to declining aquifer coupled with economic downside risks emanating from possible upswing in interest rate and input costs, the study analyzes (1) long-term financial viability of cotton-based cropping systems adopting cover crops and tillage systems as soil health practices while considering producers with varying risk preferences under different irrigation rates; and (2) economic impact of interest rates on variable cost and gross margin under different cotton-based cropping systems (i.e., continuous cotton-conventional tillage, wheat-cotton rotation, and continuous cotton-rye cover) and irrigation rates. Monte Carlo Simulation is used to calculate the different economic metrics (i.e., yield, gross margin, gross revenue, and variable cost) on a per-acre and annual basis from the experimental data of the Agricultural Complex for Advanced Research and Extension Systems (Ag-CARES) site in Lamesa, Texas. Stochastic Efficiency with Respect to a Function (SERF) is employed to rank the most preferred cropping systems based on the certainty equivalent (CE) values. Lastly, stoplight charts is used to summarize the likelihood of each scenario between the specified targets.
The results of the analysis reveals that wheat-cotton rotation is the most profitable and economically preferred system in the region when compared to continuous cotton-conventional tillage and continuous cotton-rye cover in terms of yield, gross margin, and variable cost. On the other hand, for gross revenue, continuous cotton-rye cover outperforms a wheat-cotton rotation and continuous cotton-conventional tillage and appears to be the most economically viable option. Moreover, across all scenarios and economic metrics, high irrigation rate exhibits sound economic performance as shown by high potential risk-adjusted outcome or CE values for yield, gross margin, and gross revenue while low CE values for variable cost under different risk aversion levels. As to the computed risk premiums, individual producer would be better off by switching from continuous cotton- conventional tillage to wheat-cotton rotation as if offers higher productivity benefits in terms of yield (i.e., 110.65 lb/acre to 135.24 lb/acre for low irrigation rate, 189.30 lb/acre to 256.94 lb/acre for medium irrigation rate, and 154.84 lb/acre to 205.82 lb/acre for high irrigation rate) and gross margin (i.e., 182.19/acre for low irrigation rate, 175.88/acre for medium irrigation rate, and 114.64/acre for high irrigation rate), while lower risk premiums for variable cost ranges from -185.09/acre for low irrigation rate, -204.95/acre for medium irrigation rate, and -225.02/acre for high irrigation rate . With regard to gross revenue, continuous cotton-rye cover is considered to be the most economically beneficial system for producers with risk premiums ranging from 44.64/acre for low irrigation rate, 44.62/acre for medium irrigation rate, and 49.57/acre for high irrigation rate. The study also employs scenario analysis for multiple interest rates (i.e., 4%, 6%, 8%, 10%, 12%, and 14%). Across these scenarios, wheat-cotton rotation displays superior profitability performance compared to other systems as shown by lower CE values for variable cost and higher CE values for gross margin across different irrigation applications. Moreover, wheat-cotton rotation offers lower probability of losses across all interest rate scenarios and irrigation rates for variable cost and gross margin.
Overall, this analysis offers new insights on the vital role of soil heath management practices (SHMP) in boosting financial viability of cotton-based cropping systems. Given this, wider adoption of SHMP across all stakeholders especially agricultural producers shall be promoted as a risk management approach to uplift the financial gain and capabilities in this sector
Go Back to Rest: Coffin Hardware Analysis at the Old Canaan Cemetery, Harrison County, Texas
No full textThe Old Canaan Cemetery, located in Harrison County, Texas, was excavated by Texas Tech faculty and students in 2021, 2023, and 2024. The project was initiated by the community's descendants in 2018. The cemetery was likely established in 1870 alongside the Old Canaan Missionary Baptist Church but was separated from the church due to racial intimidation tactics in the late 1939. During the excavations, 13 burials were uncovered, and an estimated 121 burials were identified. This thesis focuses on the coffin hardware from the excavated burials to establish a chronology, explore socioeconomic status by identifying stylistic lag periods, and evaluate racial inequality through these periods. Based on the coffin hardware analysis, the burials occurred between the 1900s and the 1930s. Seven of the 13 burials contained significant stylistic lag, with females and young adults having a higher prevalence of significant stylistic lag, indicating lower socioeconomic status. When compared to two African American cemeteries and two White cemeteries, the Old Canaan Cemetery displayed a higher prevalence of significant stylistic lag than the White cemeteries. The coffin hardware from the African American cemeteries displayed more significant stylistic lag than the hardware from the White cemeteries, revealing racial inequality influenced socioeconomic status
Strategies for covalent stabilization of F1-ATPase to enhance its usability as a nanomotor.
No full textF1-ATPase, the soluble rotary motor component of ATP synthase, has been harnessed in hybrid nanodevices to convert ATP hydrolysis into mechanical motion for applications in targeted drug delivery, energy harvesting, and biosensing. The enzyme comprises a catalytic stator ring of alternating α and β subunits (α3β3), surrounding a central γ subunit that rotates during catalysis. However, its intrinsic oligomeric instability under non-physiological conditions limits functional stability, performance, and integration into devices. To address this, we developed a strategy to covalently stabilize F1-ATPase via structure-guided disulfide crosslinking at key subunit interfaces.
To guide disulfide placement, an initial mutant αQ399C/βQ441C was constructed at the C-terminal domain, but its crosslinking rendered the enzyme inactive, suggesting that this region is essential for catalytic dynamics. We therefore redirected our strategy toward the N-terminal region, where we engineered two principal variants: MGO58, engineered with disulfide bonds at both the catalytic (βV8C/αQ49C) and non-catalytic (αE84C/βQ45C) interfaces to rigidify the α3β3 stator ring; and MK59, which combines these stator crosslinks with a rotor-stator tether (αP280C/γA285C) to fully lock the enzyme assembly. Auxiliary mutants MGO55 (βV8C/αQ49C), MGO56 (αE84C/βQ45C), WTαγ (αP280C/γA285C) were constructed and used to dissect individual contributions. Engineered variants were characterized via temperature-resolved ATPase assays to evaluate catalytic efficiency across thermal gradients, nanoDSF to quantify global thermal stability and unfolding transitions, and redox-sensitive Western blotting to confirm disulfide bond formation kinetics and subunit crosslinking.
Compared to native F1-ATPase, MGO58 showed a 20 kJ/mol increase in activation energy (Ea) and a 51.5 kJ/mol gain in thermal deactivation energy (Ed), consistent with enhanced rigidity and resistance to thermal denaturation. These effects translated into a 3 °C increase in optimal catalytic temperature (Topt) and a 4 °C increase in melting temperature (Tm). Non-reducing Western blot confirmed that dual-interface disulfide tethering efficiently captures the full α3β3 stator subcomplex. In contrast, MK59 exhibited reduced thermal robustness, suggesting that excessive rigidification via rotor-stator crosslinking may impair essential conformational flexibility. Hence, the stabilized MGO58 variant advances the feasibility of F1-powered nanomotors for targeted drug delivery, biosensing, and micro-energy harvesting applications.
Altogether, this work demonstrates that strategic disulfide engineering enables fine-tuning of conformational resilience in multimeric enzymes, offering a blueprint for stabilizing F1-ATPase nanomotors for future biomedical and synthetic applications
Anti-Rotation Load Leg Effects on Crash Injury Metrics Under Child Restraint System Misuse
No full textObjective: Misuse (i.e., improper installation) of child restraint systems (CRS) remains alarmingly prevalent, with estimates ranging from 49% to 95%. Common misuse errors for rear-facing infant car seats include loose CRS installation in the vehicle and inadequate tightening of the child harness. These misuse errors compromise protection during motor vehicle crashes and intensify the risk of severe or fatal head and neck injuries (e.g., whiplash, traumatic brain, spinal, neck, chest contusions, other internal organs). Although anti-rotational devices, such as load legs, have shown potential in mitigating injury metrics, their efficacy for rear-facing infant CRS misuse conditions and condition combinations remains underexplored. Therefore, this research aimed to evaluate the use of load legs on head and neck injury metrics under two common misuse conditions: loose installations and loose harnesses.
Methods: Dynamic frontal sled tests were conducted using a 12-month-old instrumented CRABI anthropomorphic test device (ATD). The ATD was secured in a rear-facing infant car seat, installed using the flexible lower anchor tether system, and tested on two distinct sled configurations: the Federal Motor Vehicle Safety Standard (FMVSS213) bench and the Consumer Reports CR35/35 bench, which replicates a more severe crash pulse. A full factorial 23 design of experiments (DOE) with no replication was executed to assess the effects of three categorical factors: load leg use, seat installation tightness, and harness tightness, yielding eight unique configurations per test protocol. Injury metrics analyzed included HIC36, HIC15, 3ms head clip, 3ms chest clip, neck resultant force, and neck resultant moment. Least squares regression modeling and ANOVA were used to assess main effects and trends across both protocols.
Results: Across both test protocols, the load leg repeatedly emerged as a dominant performer in the reduction of head injury metrics, with effectiveness and tradeoffs varying by crash severity and installation quality. Under FMVSS213 (30 mph), a full factorial ANOVA (R20.84) with a percentage comparison of least squared means showed significant reduction in HIC36 by 16.3% (p=0.0253) but increased the 3ms chest clip by 12.9% (p=0.0059). HIC15, 3ms head clip, neck resultant force and neck resultant moment results were mixed, but none reached statistical significance (p>0.05). Under CR 35/35 (35mph), model fits remained strong (R20.86) and the load leg drove larger least squared means percentage reductions; HIC36 by 25.8% (p=0.0247), HIC15 by 27.8% (p=0.0174) and 3ms head clip by 14.1% (p=0.0185). Neck resultant force increased and 3ms chest clip and neck moment decreased, however these effects were not statistically significant. Kinematic overlays confirmed that the load leg reduced seat back rotation by 13o under proper installation and by 6o-19o across misuse cases. Head contact with the blocker plate was eliminated with load leg use across all condition except for the combined seat and harness misuse condition. In comparing FMVSS213 with CR35/35 protocols using eta squared (2), shows the 3ms chest clip effect size greatest under FMVSS213, whereas head and neck 2 metrics become more dominant under the CR35/35, the higher severity test.
Conclusion: This study demonstrates that load legs improve CRS performance by reducing rotation and injury metrics under both ideal and misuse scenarios, though tradeoffs exist for mixed injury metrics. Their efficacy is more pronounced under high-energy crash conditions (CR35/35) for head and neck metrics and under lower severity crash conditions (FMVSS213) for thoracic injury. These findings validate the role of anti-rotation features as effective safety enhancements and highlight the importance of incorporating misuse-tolerant design features and user education into CRS safety strategies
Optimization of freeze-dried exudate additive for ground beef palatability and quality attributes
No full textThe objective of this study was to evaluate the shelf-life, oxidative stability, and palatability attributes of 0%, 0.5%, 1,% and 2% freeze-dried exudate (FDE) across 2 lean points of 80% and 93% lean ground beef (GB). Sixteen 4.5-kg chubs of GB at each lean point were designated to different 3.6-kg batches of FDE percentage by weight at 0%, 0.5%, 1% and 2% FDE (N=4). Within each batch, four repetitions (n=16) of five 151-g patties were formed and designated for 5 different analyses. Three patties were used for a 96 h retail display and sampled at 0, 48, and 96 h for microbial enumeration and determination of thiobarbituric acid reactive substances (TBARS). Instrumental and descriptive color of the 96-h samples were evaluated every 12 h. The remaining two patties were designated for cooking for descriptive sensory analysis or volatile compound analysis. Descriptive color scores had a %FDE by time interaction (P0.05). There was an interaction of %FDE overtime for a* values (P0.05) but the control had greater lipid oxidation than 1% FDE at 48 h (P<0.05). Descriptive sensory attributes primarily differed by lean point (P≤0.012) and only cohesiveness, hardness, brown/roasted, and burnt differed with %FDE (P≤0.008) where they were the greatest with 2% FDE inclusion (P<0.05). Beef-identity also had a %FDE by lean point interaction (P=0.041) where FDE inclusion in 80% lean increased intensity ratings (P<0.05). Freeze-dried exudate inclusion at 2% increased the concentration of Maillard derived volatiles primarily consisting of pyrazines, Strecker aldehydes, and sulfides (P<0.05). Partial least squares regression of key volatile compounds as explanatory variables and sensory attributes as dependent variables found an association of these Maillard compounds with increased flavor attribute ratings of brown/roasted and beef-identity with loadings towards 2% FDE inclusion in 80% lean. Freeze-dried exudate demonstrated an ability to alter ground beef flavor development, but this came at the detriment of discoloration and decreased shelf-life. Further optimization may allow for the utilization of exudate as a natural ingredient for ground beef and may also have viability in other products