Engineering Commons^™

Ai-Driven Traffic Scene Understanding Using Static Lidar Sensors, Elham Binshaflout, Chaima Zaghouani, Charalampos Antoniadis, Hakim Ghazzai, Nawfal Guefrachi, Ahmad Alsharoa, Sameh Najeh, Gianluca Setti

Electrical and Computer Engineering Faculty Research & Creative Works

Traffic congestion and road safety remain critical challenges in urban environments, driving the need for more effective traffic monitoring solutions. While recent advancements in computer vision have enhanced traffic perception, the dynamic viewpoint of autonomous vehicles is often insufficient for comprehensive traffic management. To address this gap, we propose an AI-driven framework for enhanced traffic scene understanding using static LiDAR sensors at road intersections. The system collects 3D point clouds from roadside static LiDAR sensors, providing a complete view of vehicles and pedestrians. We integrate state-of-the-art 3D object detection (i.e., PV-RCNN) and instance segmentation models (i.e., PointGroup3heads) to accurately identify …

Efficient Decoupling Capacitor Impact Calculation, Faye Squires, Yifan Ding, Matthew Doyle, Matteo Cocchini, Samuel Connor, Francesco De Paulis, Albert E. Ruehli, Chulsoon Hwang, Lijun Jiang

Electrical and Computer Engineering Faculty Research & Creative Works

Methods of optimizing decoupling capacitor placement on power distribution networks (PDNs) are often limited due to the computational complexity required to calculate the impact of connecting loads to an impedance matrix with hundreds of rows and columns. This work proposes that by removing all but one member of the impedance matrix before calculating, checking the impact of adding capacitors to the matrix can be done efficiently, and optimization methods can be viable even when requiring millions of impedance calculations.

Usb 3.0 Ibis-Ami Model Construction Using Measurement And Neural Network, Jiahuan Huang, Wenchang Huang, Muqi Ouyang, Hank Lin, Bin Chyi Tseng, Chulsoon Hwang

Electrical and Computer Engineering Faculty Research & Creative Works

The input/ output modes are essential for high-speed signal integrity analysis and channel simulation. This work aims to develop a method for generating an IBIS-AMI model for USB 3.0 using measurement data. Instead of requiring a specially designed motherboard with test points for specific measurements, this method uses measurement data obtained from an assembled motherboard. The only available data for measurement in this case is the output voltage waveform from the USB 3.0 port on the motherboard. To address this, a novel approach is proposed to extract all the required parameters for the IBIS-AMI model from a single available measurement …

Optimized Modeling Of Pcb Vias With Nonfunctional Pads And High-Frequency Behavior Up To 150 Ghz, Mehdi Mousavi, Kevin Cai, Chaofeng Li, Sathvika Bandi, Manish Mathew, Mehdi Khaleghi, Shameem Ahmed, Donghyun Kim

Electrical and Computer Engineering Faculty Research & Creative Works

This paper presents an enhanced closed-form approach for modeling and optimizing high-frequency PCB vias, implemented in Python and validated against industry standard tools such as ADS and HFSS. The model incorporates resistance alongside inductance and capacitance to capture frequency-dependent losses and integrates non-functional pads (NFPs), demonstrating significant improvements in signal integrity by reducing reflections and enhancing return loss, particularly at 100 GHz. The methodology extends the frequency range of previous models from 100 GHz to 150 GHz, ensuring compatibility with next-generation standards like PCIe Gen 6. Validation results show insertion loss deviations under 3 dB and consistent return loss across …

Modeling Of Slot-Backed Microstrip Line For Emi Applications, Jongsuk Hyun, Wenchang Huang, Chulsoon Hwang

Electrical and Computer Engineering Faculty Research & Creative Works

Radiated emissions from the noise-generating components in modern electronic devices are a significant concern for both electromagnetic interference and RF interference. Shielding cans are commonly used to suppress emissions from noise sources, but accurate shielding effectiveness evaluation requires a clean, well-defined radiation source that can reliably mimic real emissions for repeatable measurements. Slot-backed microstrip antennas provide a practical alternative to loop antennas, offering zero height, low parasitic radiation, and seamless printed circuit board integration. This article proposes an analytical model for estimating the magnetic dipole moment of slot-backed microstrip structures. The model captures both the discontinuity effects and radiated characteristics …

Design Strategies For Skew Compensation In Highspeed Pcb Strip Line Interconnects, Sathvika Bandi, Reza Asadi, Zhekun Peng, Srinivas Venkataraman, Granthana Rangaswamy, Santosh Pappu, Xu Wang, Donghyun Kim

Electrical and Computer Engineering Faculty Research & Creative Works

This paper presents a comprehensive analysis of the impact of intra-pair PN skew compensation in printed circuit board (PCB) strip line (SL) traces, for a high-speed 224 Gbps lane for the first time. The study investigates the effects of skew compensation placement both with and without via discontinuities. Detailed evaluations are performed in both time and frequency domains, examining critical parameters such as time-domain reflectometry (TDR), input impedance, return loss, insertion loss, and common-mode S -parameters. The findings reveal that, in a simple strip line trace without via discontinuities, the location of skew compensation has negligible influence on signal margins. …

Enhanced Continual Reinforcement Learning-Based Output Feedback Control Of Heterogeneous Quadrotors Formation, Ehsan Soleimani, S. Jagannathan

Electrical and Computer Engineering Faculty Research & Creative Works

This paper presents a unified framework for the safe and optimal control of heterogeneous quadrotor unmanned aerial vehicles (QUAVs) in formation, enabling multitask missions without requiring precise system dynamics. To address partial state observability, a multilayer neural network (MNN) observer is designed to estimate unmeasured states. Reinforcement learning (RL) is employed for optimal control utilizing an MNN ensuring adaptability. Barrier Lyapunov Functions (BLFs) are integrated into the RL framework to enforce safety by maintaining QUAVs within predefined constraints. An enhanced continual learning (ECL) method is proposed to improve the adaptability of MNNs. This method enables effective multitask learning while mitigating …

Topology And Parameter Joint Identification In Imbalanced Low-Voltage Distribution Networks Based On Load Characteristic Propagation, Yanan Zhang, Gan Zhou, Huan Mao, Wei Gu, Yanjun Feng, Rui Bo

Electrical and Computer Engineering Faculty Research & Creative Works

Low-voltage distribution networks often suffer from incomplete or outdated network records, making it challenging to obtain the topology and line parameters under actual operating conditions. To address this issue, a joint identification method is proposed based on the propagation of load transient characteristics. First, the principle of load characteristic propagation is elaborated, and the concept of coupling impedance is introduced. Second, a set of linear regression equations is established based on the changes in current and voltage of the terminal measurements before and after load switching, and then these equations are solved using the least squares method to form the …

Benchmarking Deep Learning Architectures For Ecg-Based Multi-Label Heart Disease Prediction Using Mimic-Iv Database, Eyiara Oladipo, Sarwar Nazrul, Mohamed Nafea

Electrical and Computer Engineering Faculty Research & Creative Works

Cardiovascular disease (CVD) is a leading cause of global mortality, accounting for an estimated 17.9 million deaths annually. CVD is broadly defined as a group of medical conditions influenced by modifiable or non-modifiable risk factors that affect the heart's ability to function properly. Machine learning (ML) has emerged as a powerful tool for analyzing complex medical data, aiding in early detection and accurate diagnosis of CVD and improving patient outcomes. Recent studies proposed various deep learning (DL) architectures for detecting CVD, yet there is a lack of robust benchmarks for comparing their performance on large-scale databases. In this work, we …

Structural Analysis Of Multi-Domain Dynamic Systems Modeled Via Bond Graphs, Arnold A. Fernandes, Jonathan W. Kimball

Electrical and Computer Engineering Faculty Research & Creative Works

Structural Observability (SO) and Structural Monitorability (SM) are structural properties utilized to determine the state and fault-free operation of components, respectively, in a bond graph (BG) model. BGs enable qualitative system analysis, evaluating whether existing sets of sensors and actuators ensure Structural Observability (SO) and Structural Controllability (SC) without knowledge of parametric values. Furthermore, the analysis determines whether there are sufficient sensors available to identify component faults accurately. This work provides a framework for automated sensor placement in a multi-domain physical system while analyzing the SO and SM properties. The MATLAB Structural Analysis Toolbox (MATSAT) conducts sensor placement in a …

Honey-Reram Enabled Sustainable Edge Ai System For Iot Applications, Jinhui Wang, Feng Zhao, Mohammad Rafeeq Khan, Md Mehedi Hasan Tanim, Zoe Templin, Harshvardhan Uppaluru

Electrical and Computer Engineering Faculty Research & Creative Works

This paper is toward a promising solution to address the environmental sustainability challenge in computing by building brain-inspired and green non-Von Neumann systems with Resistive Random-Access Memory (ReRAM) made from natural organic materials, honey, for energy-efficient operation, renewable material resources, sustainable device manufacturing, and environmentally-friendly disposal. In this paper, honey-ReRAM and its arrays are firstly manufactured and tested. The resistance modulation mechanism of honey-ReRAM is analyzed and investigated. Then a Computing-in-Memory (CIM) architecture based on honey-ReRAM for edge AI and IoT applications is proposed and evaluated. The experimental results indicate that the proposed edge AI systems with the VGG8 and …

Centralized And Federated Heart Disease Classification Using Uci Dataset: A Benchmark With Interpretability Analysis, Mario Padilla Rodriguez, Eyiara Oladipo, Mohamed Nafea

Electrical and Computer Engineering Faculty Research & Creative Works

Cardiovascular disease (CVD) is a leading cause of global mortality, highlighting the need for accurate diagnostic methods. This study benchmarks centralized and federated learning (FL) algorithms for heart disease binary classification using the UCI dataset, which includes 920 patient records from four hospitals in the USA, Hungary, and Switzerland. Our benchmark is supported by Shapley-value as well as Local Interpretable Model-agnostic Explanations (LIME) interpretability analyses to quantify feature importance for classification. In the centralized setup, various classification algorithms are trained on pooled data, with the Naive Bayes classifier achieving the highest test accuracy of 81.1%. Further, FL algorithms with four …

Corona Discharge To Touchscreen Modeling Using Nonlinear Time-Dependent Corona Streamer Propagation Model In Spice, Zhekun Peng, Daniel Szanto, Jianchi Zhou, Darryl Kostka, David Pommerenke, Daryl G. Beetner

Electrical and Computer Engineering Faculty Research & Creative Works

SPICE-based methods for predicting coupling from an ESD-induced corona streamer to printed-circuit board (PCB) structures beneath a touchscreen display are evaluated in this paper. Results demonstrate that the non-linear time-dependent propagation model can capture the coupling much better than other models and accurately predict the overall current waveform.

Microwave Photonics-Assisted Interrogation Of Fiber-Optic Interferometric Sensors With Joint Frequency-Time Domain Analysis, Ruimin Jie, Jie Huang, Chen Zhu

Electrical and Computer Engineering Faculty Research & Creative Works

Fiber-optic interferometers are widely used in localized sensing applications due to their compact size, high sensitivity, and immunity to electromagnetic interference. In this paper, we propose and experimentally demonstrate a novel interrogation scheme for fiber-optic interferometric sensors, utilizing microwave photonics (MWP) and joint frequency-time domain analysis. As a proof of concept, a miniature fiber in-line Fabry-Perot interferometer (FPI) is integrated with a microwave photonic single-passband filter, enhanced by a dispersion compensation module to improve sensing performance. By applying an inverse Fourier transform to the system's complex frequency response, the time-domain representation of the signal is obtained, translating spectral shifts of …

Miniaturized Wearable Biosensors For Continuous Health Monitoring Fabricated Using The Femtosecond Laser-Induced Graphene Surface And Encapsulated Traces And Electrodes, Homayoon Soleimani Dinani, Tatianna Reinbolt, Bohong Zhang, Ganggang Zhao, Rex E. Gerald, Zheng Yan, Jie Huang

Electrical and Computer Engineering Faculty Research & Creative Works

Wearable sensors are increasingly being used as biosensors for health monitoring. Current wearable devices are large, heavy, invasive, skin irritants, or not continuous. Miniaturization was chosen to address these issues, using a femtosecond laser-conversion technique to fabricate miniaturized laser-induced graphene (LIG) sensor arrays on and encapsulated within a polyimide substrate. The femtosecond laser-converted conductive traces can have a size of 20 to 2 μm compared to the traditionally larger CO2 laser dimensions of around 300 to 100 μm. This marks a 93-98% decrease in trace size when using a femtosecond laser. This miniaturization allows for the ability to process temperature, …

S²Il: Structurally Stable Incremental Learning, S. Balasubramanian, P. Yedu Krishna, Talasu Sai Sriram, M. Sai Subramaniam, Manepalli Pranav Phanindra Sai, Ravi Mukkamala

Computer Science Faculty Publications

Feature Distillation (FD) strategies are proven to be effective in mitigating Catastrophic Forgetting (CF) seen in Class Incremental Learning (CIL). However, current FD approaches enforce strict alignment of feature magnitudes and directions across incremental steps, limiting the model’s ability to adapt to new knowledge. In this paper, we propose Structurally Stable Incremental Learning (S²IL), a FD method for CIL that mitigates forgetting by focusing on preserving the overall spatial patterns of features which promote flexible (plasticity) yet stable representations that preserve old knowledge (stability). We also demonstrate that our proposed method S²IL achieves strong incremental accuracy and outperforms other FD …

Energy-Based Deep Incomplete Multi-View Clustering, Ziyu Wang, Yiming Du, Rui Ning, Lusi Li

Computer Science Faculty Publications

Incomplete multi-view clustering (IMVC) deals with real-world scenarios where certain views are partially missing, posing significant challenges to effective clustering. Most existing IMVC approaches face a trade-off: imputation-free methods suffer from information bias and imbalance, while full-imputation methods risk introducing and propagating noise. To overcome these limitations, we propose Energy-Based Deep Incomplete Multi-View Clustering (Energy-DIMC), a novel selective-imputation framework that leverages energy-based models (EBMs) to guide reliable imputations and robust clustering. EBMs assess data compatibility by assigning lower energy to more coherent structures, effectively modeling complex inter-view and inter-sample dependencies. Inspired by EBMs, Energy-DIMC integrates four key components: 1) a …

Contextual Memory Recall: A Novel Metric For Class Incremental Learning, Balasubramanian S, Sai Subramaniam M., Sai Sriram Talasu, Yedu Krishna P., Pranav Phanindra Sai M., Darshan Gera, Ravi Mukkamala

Computer Science Faculty Publications

We propose a novel metric for class incremental learning (CIL) called Contextual Memory Recall (CMR), which evaluates how well a CIL model recalls previously learned classes when given relevant past cues. Inspired by human memory, CMR offers newer insights into continual aspects of a CIL model that were not addressed by previously proposed metrics for CIL. Specifically, the standard metric, average incremental accuracy (AIA), overlooks the quality of evolving feature representations, whereas our proposed CMR accounts for it. As a result, methods using feature distillation perform well under AIA but poorly under CMR, while those without feature distillation excel under …

Nanomaterial-Enabled Enhancements In Thylakoid-Based Biofuel Cells, Amit Sarode, Gymama Slaughter

Center for Bioelectronics Publications

Thylakoid-based photosynthetic biofuel cells (TBFCs) harness the inherent light-driven electron transfer pathways of photosynthesis to enable sustainable solar-to-electrical energy conversion. While TBFCs offer a unique route toward biohybrid energy systems, their practical deployment is hindered by sluggish electron transfer kinetics, unstable redox mediators, and inefficient interfacing between biological and electrode components. This review critically examines recent advances in TBFCs, with a focus on three key surface engineering strategies: (i) incorporation of nanostructured materials to enhance electrode conductivity and surface area; (ii) application of redox mediators to facilitate charge transfer between photosynthetic proteins and electrodes; and (iii) functional exploitation of individual …

Emi Mitigation, Material Characterization, Transformer Equivalent Circuit Modeling, Reza Vahdani

Masters Theses

Modern high-frequency electronic systems demand precise characterization and modeling techniques to ensure signal integrity and electromagnetic compatibility. This thesis presents three core studies focused on real-world challenges in high-speed and power electronics: EMI mitigation using 3D printed absorbers, wideband liquid dielectric characterization, and accurate transformer modeling.

The first study demonstrates a targeted approach to mitigating electromagnetic interference (EMI) in a commercial router. By using holography imaging to identify radiation hotspots, custom absorber structures were designed with commercially available materials, fabricated via 3D printing, and applied directly to emission sources. Radiated emission tests in a reverberation chamber showed up to 9 …

Frequency-Tracker And Power Supply For Piezoelectric Lunar Dust Removal Actuator, Praneeth Uddarraju

Masters Theses

"This work presents a reconfigurable phase-locked loop (PLL)-based control system for the resonant excitation of piezoelectric actuators aimed at automated removal of particulate contaminants from photovoltaic (PV) surfaces—particularly in extraterrestrial environments such as the lunar surface. Regolith or lunar dust buildup on solar panels is a serious hazard to the effectiveness of energy harvesting on extended missions. By using high-frequency structural excitation and inertial forces the suggested system removes surface impurities. Tunable Sallen-Key low-pass filters for reliable feedback conditioning are used in conjunction with a digitally implemented PLL on an FPGA (XLR8 platform) to precisely lock the drive frequency to …

Lifelong Machine Learning With Adaptive Resonance Theory, Sasha Petrenko

Doctoral Dissertations

"This publication option dissertation is composed of three papers concerning the study of the problem lifelong machine learning with Adaptive Resonance Theory (ART) algorithms. Lifelong learning (L2) is a challenging machine learning paradigm that both encompasses and formalizes the fields of continual learning and incremental learning. The field is concerned with the mitigation of the phenomenon of catastrophic forgetting whereby learning agents that are faced with incrementally novel information deleteriously overwrite previous knowledge if that learning process is not regularized to counteract this consequence. ART algorithms solve this stability-plasticity dilemma by optimally assigning learning to categories or instantiating new knowledge …

Systematic Esd Analysis And Modeling For Electronic Device, Zhekun Peng

Doctoral Dissertations

Systematic ESD analysis provides good pre-compliance to ESD robustness evaluation on the electronic device from device level to component/system level to on-chip level. The whole process involves corona discharge on display, system level ESD analysis on PCB for race condition and transient response and 3D IC package impact to on-die ESD.

ESD to the display cover glass can damage touchscreen traces by sparkless corona discharges on the glass surface. A non-linear time dependent transmission-line model is proposed to model corona streamer propagation in terms of the coupling current and propagation speed. Results are highly promising to model the corona discharge …

Measurement Method And Applications Of Transfer Function In Rf Desensitization Problem, Xiangrui Su

Doctoral Dissertations

Radio frequency (RF) desensitization issues comprise two components: noise radiation sources and the transfer function from noise sources to the victim antenna. RFI is a critical challenge in modern electronic systems, particularly in densely packed environments. This work presents a comprehensive study of RFI, addressing key aspects through three novel contributions. First, a transfer function measurement method is developed for compact metallic enclosures. This method provides a precise characterization of the electromagnetic (EM) environment within confined spaces, enabling accurate identification of interference pathways. Second, an EM emission management analysis framework is proposed, leveraging transfer functions to quantify and mitigate interference …

Enhancing Performance Of Lstm For Water Level Forecasting In Lower Mekong River With Feature Selection Strategy, Chanreng Sey Nhim

Chulalongkorn University Theses and Dissertations (Chula ETD)

This study addresses key research gaps, including the lack of systematic feature selection, structured hyperparameter optimization, and comprehensive long-term and cross-location performance evaluation in the Lower Mekong River Basin to enhance the water level forecasting performance. The research has utilized a Vanilla Long Short-Term Memory (LSTM) network trained on 22 years of historical water level data from the Mekong River Commission (MRC) and 16 localization meteorological features from Open-Meteo. The methodology uses Grid Search optimization for hyperparameter tuning and employs the Spearman rank correlation coefficient as a filter feature selection technique to identify relevant input variables and minimize redundancy. Performance …

A Study On The Impact Of Superconducting Fault Current Limiteron Arc Flash Hazard In An Industrial System, Kittipong Anantanasap

Chulalongkorn University Theses and Dissertations (Chula ETD)

The increasing integration of distributed generation raises short-circuit current levels, which lead to concerns about equipment overheating, insulation deterioration, and system stability issues. The Resistive Superconducting Fault Current Limiter (R-SCFCL) offers an effective solution by dynamically increasing circuit impedance during faults without affecting normal operation. However, its integration may alter the fault-clearing time of the protection system, which is crucial for arc-flash safety. This study evaluates the impact of R-SCFCL integration on arc-flash hazards in an industrial power system, utilizing DIgSILENT PowerFactory and calculations based on IEEE 1584. Results indicate that while the R-SCFCL effectively reduces arcing current, the incident …

Vehicle Routing Problem With Time Constraint From Multiple Origins To Selective Destinations By Using Genetic Algorithm, Manocha Kruetet

Chulalongkorn University Theses and Dissertations (Chula ETD)

This thesis addresses the vehicle routing problem with time constraints from multiple origins to selective destinations by using a genetic algorithm for efficient path planning. The focus is on optimizing field collection and repossession activities in the financial industry to enhance operational efficiency and customer satisfaction. The proposed approach integrates two main components: a tailored genetic algorithm for multi-objective optimization and a graph neural network for precise travel time estimation. The genetic algorithm is designed to minimize total travel distance or travel time and balance the workload among field collectors, while the graph neural network provides travel time estimates to …

Performance Evaluation Of Strategic Bandwidth Block Assignment For Interference Reduction In High-Density 5g Nr Cell Environments, Pasapong Wongprasert

Chulalongkorn University Theses and Dissertations (Chula ETD)

Due to the increasing demand for mobile data, mobile network operators (MNOs) typically rely on cell densification to increase capacity. However, doing so also introduces more inter-cell interference (ICI), leading to lower spectral efficiency and making further capacity enhancements challenging. This thesis explores a bandwidth block assignment method for ICI mitigation. The core idea is to divide the bandwidth of an NR carrier into two partially overlapping bandwidth parts (BWPs) to reduce the impact of ICI. By allowing the two BWPs to share the same synchronization signal block (SSB), the same attach and handover configurations can be applied. A modified …

Real-Time Monitoring System For Sedentary Behavior Using Pose-Estimation Approach, Pheng Ou Chea

Chulalongkorn University Theses and Dissertations (Chula ETD)

Long-term sedentary behavior, defined as extended periods of sitting or inactivity, increases the risk of obesity, cardiovascular disease, and other health problems. These risks are often underestimated due to low levels of physical activity. This study proposes the use of You Only Look Once (YOLO)–based pose estimation for dwell-time analysis to monitor sedentary behavior. We examined posture over an entire day to assess prolonged inactivity and encourage individuals to balance energy expenditure with available time—such as engaging in exercise in the early morning or evening according to recommended guidelines. For employees across generations, as well as for vulnerable groups such …

Bismuth Surfactant-Mediated Epitaxy Of Gaas(Bi) On On-Axis Ge(100) Substrates, Valgeir Sigmarsson

Chulalongkorn University Theses and Dissertations (Chula ETD)

In this work, Bi-surfactant-mediated epitaxy is applied in GaAs(Bi) film deposition on nom- inally on-axis Ge(100) substrates to investigate its effect on mitigating the formation of anti-phase domains (APDs) inherent to the polar-on-non-polar GaAs/Ge(100) heterostructure. GaAs(Bi) layers were deposited by solid-source molecular beam epitaxy (MBE) with a V/III ratio of ~90 and stepwise variations in Bi flux between samples. The Bi-free GaAs sample exhibited an unstable RHEED pattern indicating ill-ordered growth mode, confirmed by AFM showing an amor- phous surface with an RMS value of ~20 nm. In contrast, Bi-assisted growth produced stable RHEED (1×3) surface reconstructions. AFM revealed single-crystal …