Selected work

Representative Papers

Real time fault detection in 3D printers using Convolutional Neural Networks and acoustic signals

Feb 17, 2026

This study addresses the challenge of low-cost, real-time detection of mechanical faults in fused deposition modeling (FDM) 3D printing—such as nozzle clogging and filament breakage—which significantly compromise print quality and system reliability. To this end, the authors propose a non-intrusive monitoring approach that leverages acoustic signals and a convolutional neural network (CNN) to classify multiple common printing anomalies in real time. By collecting operational audio from the printer and constructing a dedicated fault dataset, the method eliminates the need for expensive sensors or manual intervention. Experimental results demonstrate that the proposed framework achieves high classification accuracy and robustness across diverse fault scenarios, offering an efficient, scalable, and cost-effective solution for real-time process monitoring in 3D printing.

0 citationsRead paper

Diffusion-Driven Deceptive Patches: Adversarial Manipulation and Forensic Detection in Facial Identity Verification

Jan 14, 2026

This work addresses the vulnerability of facial biometric systems to adversarial attacks by proposing a highly realistic and interpretable adversarial patch generation method. By integrating diffusion models with the Fast Gradient Sign Method (FGSM), the approach simultaneously optimizes adversarial perturbations while performing luminance correction and Gaussian smoothing, achieving high visual fidelity (SSIM of 0.95). The framework innovatively incorporates a ViT-GPT2 architecture to translate identity information into semantic descriptions, thereby enhancing forensic interpretability. Furthermore, it systematically evaluates the robustness of recognition models under adversarial conditions through a fusion of perceptual hashing and image segmentation techniques. This end-to-end pipeline balances high attack efficacy with strong interpretability, offering a novel direction for improving the security of biometric authentication systems.

0 citationsRead paper

Modeling Policy and Resource Dynamics in the Construction Sector of Developing Countries: A System Dynamics Approach Using Sudan as a Case Study

Jan 01, 2026

This study addresses chronic challenges in the construction sector of developing countries—namely project delays, cost overruns, and inefficient regulation—by proposing an integrated modeling framework that combines system dynamics with a genetic algorithm. Leveraging empirical data from Sudan and expert knowledge, the model simulates the interplay among labor, materials, financing, and policy implementation delays in infrastructure projects, supported by multi-scenario simulations and sensitivity analyses. The findings reveal that streamlining regulatory processes can reduce project delays by 32%, while enhanced investment in human capital lowers cost overruns by 28%. In contrast, interventions targeting only material or financial supply yield limited improvements. These results offer low-income countries a high-impact intervention strategy centered on regulatory reform and human capital development.

0 citationsRead paper

Preliminary Analysis of Construction Work Zone on Roadways in Florida by Crash Severity

Jul 09, 2025

Traffic fatalities and injuries in roadway construction zones across Florida—particularly in Broward, Duval, Hillsborough, and Orange counties—are alarmingly high (averaging 71 fatalities and 309 serious injuries annually), with current safety interventions proving insufficient. To address this, we integrate the 4Es (Engineering, Enforcement, Education, Emergency Response) and 4Is (Information, Incentives, Infrastructure, Innovation) safety frameworks and develop, for the first time, a hybrid analytical system combining a multinomial logit model with machine learning–based early-warning mechanisms. This system identifies key determinants of crash severity, including construction zone type, shoulder configuration, weather and lighting conditions, presence of on-site workers, and law enforcement visibility. Results reveal several statistically significant risk factors, enabling the design of a real-time, tiered risk-alert system for drivers and transportation managers. The framework provides a scalable, evidence-based decision-support tool for intelligent construction zone safety management, offering actionable technical pathways to reduce crash-related casualties.

0 citationsRead paper

Factors Influencing Change Orders in Horizontal Construction Projects: A Comparative Analysis of Unit Price and Lump Sum Contracts

Jun 30, 2025

This study addresses the pervasive issue of horizontal change orders (COs) under the Design–Bid–Build (DBB) delivery method, which frequently trigger cost overruns and schedule delays. Leveraging historical data from 770 Florida Department of Transportation projects (581 unit-price and 189 lump-sum contracts), a discrete choice model is developed to quantitatively identify key determinants of CO frequency—including project size, duration, and work type—and to assess the differential impact of contract type on CO incidence. Its novelty lies in the first application of discrete choice modeling to inform DBB contract-type selection, moving beyond heuristic, experience-based matching. Empirical validation demonstrates that the model outperforms current practice, yielding higher accuracy in contract-type recommendation and improved precision in CO frequency prediction. The framework thus provides a practical, quantitative decision-support tool for transportation agencies to optimize procurement strategies and mitigate change-order risk.

0 citationsRead paper

Recent publications

Latest Papers

Real time fault detection in 3D printers using Convolutional Neural Networks and acoustic signals

Feb 17, 2026

This study addresses the challenge of low-cost, real-time detection of mechanical faults in fused deposition modeling (FDM) 3D printing—such as nozzle clogging and filament breakage—which significantly compromise print quality and system reliability. To this end, the authors propose a non-intrusive monitoring approach that leverages acoustic signals and a convolutional neural network (CNN) to classify multiple common printing anomalies in real time. By collecting operational audio from the printer and constructing a dedicated fault dataset, the method eliminates the need for expensive sensors or manual intervention. Experimental results demonstrate that the proposed framework achieves high classification accuracy and robustness across diverse fault scenarios, offering an efficient, scalable, and cost-effective solution for real-time process monitoring in 3D printing.

0 citationsRead paper

Diffusion-Driven Deceptive Patches: Adversarial Manipulation and Forensic Detection in Facial Identity Verification

Jan 14, 2026

This work addresses the vulnerability of facial biometric systems to adversarial attacks by proposing a highly realistic and interpretable adversarial patch generation method. By integrating diffusion models with the Fast Gradient Sign Method (FGSM), the approach simultaneously optimizes adversarial perturbations while performing luminance correction and Gaussian smoothing, achieving high visual fidelity (SSIM of 0.95). The framework innovatively incorporates a ViT-GPT2 architecture to translate identity information into semantic descriptions, thereby enhancing forensic interpretability. Furthermore, it systematically evaluates the robustness of recognition models under adversarial conditions through a fusion of perceptual hashing and image segmentation techniques. This end-to-end pipeline balances high attack efficacy with strong interpretability, offering a novel direction for improving the security of biometric authentication systems.

0 citationsRead paper

Modeling Policy and Resource Dynamics in the Construction Sector of Developing Countries: A System Dynamics Approach Using Sudan as a Case Study

Jan 01, 2026

This study addresses chronic challenges in the construction sector of developing countries—namely project delays, cost overruns, and inefficient regulation—by proposing an integrated modeling framework that combines system dynamics with a genetic algorithm. Leveraging empirical data from Sudan and expert knowledge, the model simulates the interplay among labor, materials, financing, and policy implementation delays in infrastructure projects, supported by multi-scenario simulations and sensitivity analyses. The findings reveal that streamlining regulatory processes can reduce project delays by 32%, while enhanced investment in human capital lowers cost overruns by 28%. In contrast, interventions targeting only material or financial supply yield limited improvements. These results offer low-income countries a high-impact intervention strategy centered on regulatory reform and human capital development.

0 citationsRead paper

Preliminary Analysis of Construction Work Zone on Roadways in Florida by Crash Severity

Jul 09, 2025

Traffic fatalities and injuries in roadway construction zones across Florida—particularly in Broward, Duval, Hillsborough, and Orange counties—are alarmingly high (averaging 71 fatalities and 309 serious injuries annually), with current safety interventions proving insufficient. To address this, we integrate the 4Es (Engineering, Enforcement, Education, Emergency Response) and 4Is (Information, Incentives, Infrastructure, Innovation) safety frameworks and develop, for the first time, a hybrid analytical system combining a multinomial logit model with machine learning–based early-warning mechanisms. This system identifies key determinants of crash severity, including construction zone type, shoulder configuration, weather and lighting conditions, presence of on-site workers, and law enforcement visibility. Results reveal several statistically significant risk factors, enabling the design of a real-time, tiered risk-alert system for drivers and transportation managers. The framework provides a scalable, evidence-based decision-support tool for intelligent construction zone safety management, offering actionable technical pathways to reduce crash-related casualties.

0 citationsRead paper

Factors Influencing Change Orders in Horizontal Construction Projects: A Comparative Analysis of Unit Price and Lump Sum Contracts

Jun 30, 2025

This study addresses the pervasive issue of horizontal change orders (COs) under the Design–Bid–Build (DBB) delivery method, which frequently trigger cost overruns and schedule delays. Leveraging historical data from 770 Florida Department of Transportation projects (581 unit-price and 189 lump-sum contracts), a discrete choice model is developed to quantitatively identify key determinants of CO frequency—including project size, duration, and work type—and to assess the differential impact of contract type on CO incidence. Its novelty lies in the first application of discrete choice modeling to inform DBB contract-type selection, moving beyond heuristic, experience-based matching. Empirical validation demonstrates that the model outperforms current practice, yielding higher accuracy in contract-type recommendation and improved precision in CO frequency prediction. The framework thus provides a practical, quantitative decision-support tool for transportation agencies to optimize procurement strategies and mitigate change-order risk.

0 citationsRead paper