Electrical Engineering and Computer Science

Permanent URI for this collection

https://hdl.handle.net/10315/36010

Browse

Now showing 1 - 20 of 33

Open Access
Data-driven Methods for Optimal Power Flow in Smart Grids
(2024-11-07) Wang, Junfei; Srikantha, Pirathayini
The modern power grid is undergoing significant changes, driven by increasing complexity, the integration of renewable energy sources, and the urgent need to reduce greenhouse gas emissions. These challenges necessitate more advanced methods for power grid operations. System operators continuously solve the Optimal Power Flow (OPF) problem at regular intervals to determine the most economical dispatch of power while balancing electricity supply and demand. However, traditional OPF and convex relaxation methods often face issues related to feasibility and computational speed. Recently, machine learning methods have gained considerable attention as potential solutions to these challenges. As discussed in the literature, these methods include supervised learning, hybrid approaches that combine physical solvers or equations with machine learning, and unsupervised learning. Despite these advancements, there remain research gaps that need to be addressed. In this thesis, three mechanisms for addressing the OPF problem from different perspectives are proposed. Firstly, a supervised learning algorithm with a subsequent feasibility calibration method is introduced. Secondly, a generative adversarial network (GAN) with a representation learning module is studied and employed as an optimizer for the OPF problem. Lastly, the nearly convex nature of power flow data is investigated, motivating the development of a data-driven convex relaxation approach to solve the OPF problem. This thesis makes significant contributions to the literature by ensuring the feasibility of OPF solvers through post-process algorithms with theoretical support, relaxing the assumption of having optimal solutions for training, and demonstrating high performance of data-driven OPF methods on large systems, such as the PGLIB 2000-bus system.
Open Access
Revisiting gamut expansion for color space conversion
(2024-11-07) Le, Hoang Minh; Brown, Michael S.
Cameras and image-editing software are capable of processing images in the wide-gamut ProPhoto color space, which encompasses 90\% of all visible colors. However, when images are prepared for sharing, this rich color representation is transformed and clipped to fit within the smaller-gamut standard RGB (sRGB) color space, which represents only 30\% of visible colors. Recovering the lost color information poses a challenge due to this clipping procedure. We propose three methods to address this issue. The first method proposes a deep neural network specifically trained for wide-gamut color restoration, utilizing datasets generated by a software-based camera image signal processor that produces pairs of ProPhoto and sRGB images. The second method implements a technique that incorporates a small set of color values sampled from the original ProPhoto image that is saved together with the final smaller-gamut sRGB image. The method then uses the subsampled wide-gamut color values to estimate the original ProPhoto image from the sRGB image. The third method proposes a lightweight multi-layer perceptron (MLP) trained on pairs of ground truth and clipped ProPhoto values during the gamut compression phase. The MLP is saved as metadata in the sRGB image and can later be used to predict and restore the original wide-gamut colors during the gamut expansion phase. Additionally, we have created several large-scale public datasets of wide-gamut/small-gamut image pairs to support research on color space conversion.
Open Access
Open-Gate Junction Field-Effect Transistor for Chemical and Biomolecular Analysis
(2024-11-07) Panahi, Abbas; Ghafar-Zadeh, Ebrahim; Magierowski, Sebastian
Field-effect transistors (FETs) have emerged as a transformative platform in biosensing, effectively bridging the gap between electronics and biology by measuring the intrinsic charge of bioparticles. The development of FET-based sensing techniques can be traced back to the 1970s with the advent of the ion-sensitive FET (ISFET). This innovation stemmed from Bergveld's pioneering work on metal-oxide semiconductor-based FETs, which were modified to create ISFETs. The core innovation behind ISFETs and BioFETs was the realization that replacing the metal gate of a MOSFET-like structure with an ion-sensitive membrane could transform the device into a robust biosensor with biomolecular sensing capabilities. This modification, involving the replacement of the top gate with a solution gate, enables a sensing modality where the potential of the interface between the solution and the sensing layer (typically dielectrics) can be measured. These solid-state devices leverage the modulation of charge carriers at the semiconductor-electrolyte interface, facilitating label-free detection of biological and chemical analytes. FET biosensors offer numerous advantages, including intrinsic molecular charge sensing, real-time monitoring, high sensitivity, size scalability (ranging from a few nanometers to tens of micrometers), and compatibility with standard integrated circuit (IC) technologies such as complementary metal-oxide semiconductor (CMOS). Over the past three decades, researchers have harnessed the inherent charge sensitivity of FET sensors to detect a wide range of analytes, including ions, proteins, DNA, and other biomolecules. These sensors can selectively capture target analytes by functionalizing the FET’s gate surface with specific receptors, such as antibodies or aptamers. This capability has positioned FET biosensors as crucial tools in medical diagnostics, environmental monitoring, and drug development, driving innovation and expanding the frontiers of these fields. The widespread demand for FET sensors across diverse sensing domains underscores the critical role of standard foundry-based fabrication methodologies. Efforts to refine FET sensor fabrication processes for enhanced reliability, scalability, and reproducibility are crucial for mass production and widespread adoption. Standard foundry-based fabrication techniques ensure uniformity and consistency in device performance, enabling seamless integration into diverse sensing platforms across industries. Most standard FET sensors (or ion-sensitive FETs) have been realized by CMOS platforms, which allow these sensors to be manufactured in an almost flawless foundry process that creates ICs in our cell phones. Various CMOS-based ISFET configurations were made and used for different biosensing applications. Later, graphene and carbonaceous materials were used as novel high-performance FET biosensors. However, the real breakthrough came with the efforts to standardize graphene FETs through foundry fabrication, pioneered by companies such as Cardea Bio and Graphena. This strategic initiative has not just made graphene FETs more accessible but has democratized access to them, previously hindered by prohibitive costs. This development fosters accessibility for researchers and industrial stakeholders, opening new avenues for research and innovation. By leveraging standard foundry-based fabrication approaches, FET sensors are poised to meet the escalating demand for reliable, high-performance sensing solutions. This advancement could catalyze innovation across various fields, spanning healthcare, environmental monitoring, agriculture, and beyond. The recent introduction of the open-gate junction field-effect transistor (OG-JFET) by CMC Microsystem represents a significant advancement in silicon-foundry-based FET sensor technology standardization. The OG-JFET foundry enables the mass production of chips with diverse designs on a silicon wafer, allowing for customized configurations tailored to specific application requirements. The structural innovation of the OG-JFET, achieved by removing the top gate of a p-type JFET sensor and introducing a soft material, indicates a new era of charge-sensing capabilities within a JFET electronic structure. Understanding the intricate physics of the charge sensing mechanism and its integration with microfluidics has been pivotal in unlocking the full potential of the OG-JFET. As the technology is still in its early stages of development and scale-up, user-friendly CAD tools and dedicated portable characterization systems have emerged as a cornerstone for successful adoption and integration in academic and industrial research settings. This system enhances performance and facilitates seamless integration into diverse research environments. This thesis is dedicated to tackling the challenges inherent in comprehending the charge-sensing physics of OG-JFET sensors and exploring the design and simulation of these sensors to discern the impacts of various design parameters. At the heart of this research lies developing a portable characterization system and electro-fluidics packaging schemes to streamline testing and characterization processes. A key objective of this endeavor is to standardize these processes, thereby overcoming the barrier posed by the high cost and limited accessibility of conventional characterization equipment such as probe stations and semiconductor analyzers. By providing researchers with accessible and cost-effective alternatives, this research seeks to democratize access to this FET sensor foundry technology, thus fostering innovation and discovery in chemical and biological sensing. These concerted efforts aim to accelerate the adoption and advancement of FET sensor technology, paving the way for transformative breakthroughs in scientific inquiry and technological applications.
Open Access
Transformer Differential Protection Under Geomagnetically Induced Current Conditions
(2024-11-07) Ahmadzadeh-Shooshtari, Babak; Rezaei-Zare, Afshin
Solar activities result in fluctuations in the Earth’s magnetic field, referred to as geomagnetic disturbances (GMDs). During the GMDs, geomagnetically induced currents (GICs) flow through the power transformers with grounded neutrals. The transformers subjected to the GIC experience part-cycle core saturation creating significant magnetizing current harmonics, including dominant second harmonics. The GIC-caused harmonics can pose issues for protection systems, such as transformer differential relays. The majority of real-world transformer differential relays employ harmonic blocking (HB) modules to prevent malfunctions during transformer energization leading to inrush currents containing high levels of second harmonics. Considering that the GIC-related transformer current second harmonics can be significant, the differential relays are prone to be blocked during the GMDs and remain blocked for the faults occurring in their protection zones, which may damage the transformers and threaten the system’s resiliency. This dissertation conducts in-depth time-domain analyses on the performance of differential relays, equipped with various HB logics, for internal/in-zone faults under the GIC conditions. It is revealed that regardless of the employed HB logic, such relays are unwantedly blocked for the faults with specific ranges of fault resistances in the presence of GIC, and therefore, fail to trip for the faults. To address this issue, different unblocking schemes are proposed, which can detect the GIC flow in the transformer and override the blocking signals, generated by the relay’s HB module, during the internal faults. The effectiveness of the developed schemes is extensively investigated to ensure (i) dependability for the co-existence of internal faults and GIC and (ii) security during the GMDs in the absence of faults as well as the transformer energization and external faults, especially under the current transformer (CT) saturation. Moreover, detailed real-time hardware-in-the-loop (HIL) tests verify that a commercial transformer differential relay fails to operate for terminal and transformer turn faults owing to the harmonic blocking during the GIC when the post-fault differential current fundamental component decreases below certain levels. Furthermore, it is shown that the relay equipped with a proposed unblocking scheme is effectively unblocked in real time and operates in the internal fault cases where the commercial relay cannot trip.
Open Access
CMOS Capacitive Sensor for Cellular and Molecular Monitoring
(2024-07-22) Tabrizi, Hamed Osouli; Magierowski, Sebastian; Ghafar-Zadeh, Ebrahim
This thesis focuses on the design and implementation of complementary metal-oxide-semiconductor (CMOS) based capacitive sensors for life science applications. The use of CMOS capacitive sensors has shown to be effective in a variety of applications, including chemical solvent monitoring, cellular monitoring, and DNA analysis. Despite significant advances, major challenges still exist with the current CMOS capacitive biosensing technologies including extending the dynamic range of detection, diminishing the sensitivity to remnants, and rapid high-throughput monitoring. In the second chapter, a fully integrated capacitive sensor with a wide input dynamic range (IDR) and a digital output is proposed. The design concepts and constraints, functionality, characterization, and experimental results with chemical solvents are also demonstrated in this chapter. With this novel topology, a significant increase in the IDR has been achieved which is discussed in the same chapter. Furthermore, in this chapter, we have proposed a novel calibration-free capacitive sensing technique. The proposed technique allows for uncovering the sudden changes due to the remnants as well as gradual changes due to target molecules and cells. The input dynamic range of the system is 400fF based on the post-layout simulation results. the measured resolution of the sensor is equal to 416 aF with up to 1.27 pF input offset adjustment range using the programmable bank of capacitors with a resolution of 10 fF. In the third chapter, we present a fully integrated capacitive sensor array for life science applications. This sensing device consists of an array of 16 × 16 interdigitated electrodes (IDEs) integrated with a charge-based readout and multiplexing circuitries on the same chip. This sensing device has a wide IDR of about 100 fF and a resolution of 150 aF, and the capability of temporal, spatial, and dielectric sensing. It makes it possible to develop a low-cost, calibration-free sensing platform for life science applications. In this chapter, the functionality and applicability of the proposed sensing device have been demonstrated and discussed by introducing various chemical solvents including ethanol, methanol, and pure water. The simulation and experimental results achieved in this work have taken us one step closer to a fully automated calibration-free capacitive sensing platform for high-throughput monitoring in life science applications. In the fourth chapter, the applicability of the proposed CMOS capacitive sensor for monitoring dried DNA mass has been demonstrated with experimental results. These experiments enabled us to measure the linear effect of five different concentrations with a resolution of 45 ng/μl DNA mass in ultra-pure water. With this novel application of the CMOS capacitive sensor, we can monitor the dried DNA for DNA storage monitoring purposes. Based on the results, the detection range of sub-pico mol has been achieved which is compatible with the concentrations of DNA used in DNA memory technologies. In the fifth chapter, a rapid and accurate assessment of oral cells using our CMOS capacitive sensor chips has been demonstrated. This kind of diagnostics allows for the early detection and control of periodontal and gum diseases. the experimental and simulation results demonstrate the functionality and applicability of the proposed sensor for monitoring oral cells in a small volume of 1 µl saliva samples. These results reveal that the hydrophilic adhesion of oral cells on the chip alters the capacitance of IDEs. The presented results in this chapter set a new stage for the emergence of sensing platforms for testing oral samples.
Open Access
A Unified Multiscale Encoder-Decoder Transformer for Video Segmentation
(2024-07-18) Karim, Rezaul; Wildes, Richard P.
This dissertation presents an end-to-end trainable and unified multiscale encoder-decoder transformer for dense video estimation, with a focus on segmentation. We investigate this direction by exploring unified multiscale processing throughout the processing pipeline of feature encoding, context encoding and object decoding in an encoder-decoder model. Correspondingly, we present a Multiscale Encoder-Decoder Video Transformer (MED-VT) that uses multiscale representation throughout and employs an optional input beyond video (e.g., audio), when available, for multimodal processing (MED-VT++). Multiscale representation at both encoder and decoder yields three key benefits: (i) implicit extraction of spatiotemporal features at different levels of abstraction for capturing dynamics without reliance on additional preprocessing, such as computing object proposals or optical flow, (ii) temporal consistency at encoding and (iii) coarse-to-fine detection for high-level (e.g., object) semantics to guide precise localization at decoding. Moreover, we explore temporal consistency through a transductive learning scheme that exploits many-to-label propagation across time. To demonstrate the applicability of the approach, we provide empirical evaluation of MED-VT/MEDVT++ on three unimodal video segmentation tasks: (Automatic Video Object Segmentation (AVOS), actor-action segmentation, Video Semantic Segmentation (VSS)) and a multimodal task (Audio Visual Segmentation (AVS)). Results show that the proposed architecture outperforms alternative state-of-the-art approaches on multiple benchmarks using only video (and optional audio) as input, without reliance on additional preprocessing, such as object proposals or optical flow. We also document details of the model’s internal learned representations by presenting a detailed interpretability study, encompassing both quantitative and qualitative analyses.
Open Access
Investigating and Modeling the Effects of Task and Context on Drivers' Attention
(2024-07-18) Kotseruba, Iuliia; Tsotsos, John K.
Driving, despite its widespread nature, is a demanding and inherently risky activity. Any lapse in focus, such as failing to look at the traffic signals or not noticing the actions of other road users, can lead to severe consequences. Technology for driver monitoring and assistance aims to mitigate these issues, but requires a deeper understanding of how drivers observe their surroundings to make decisions. In this dissertation, we investigate the link between where drivers look, tasks they perform, and the surrounding context. To do so, we first conduct a meta-study of the behavioral literature that documents an overwhelming importance of the top-down (task-driven) effects on gaze. Next, we survey applied research to show that most models do not necessarily make this connection and instead establish correlations between where the drivers looked and images of the scene, without explicitly considering drivers' actions and environment. Next, we annotate and analyze the four largest publicly available datasets that contain driving footage and eye-tracking data. The new annotations for task and context show that data is dominated by trivial scenarios (e.g. driving straight, standing) and help uncover problems with the typical data recording and processing pipelines that result in noisy, missing, or inaccurate data, particularly during safety-critical scenarios (e.g. intersections). For the only dataset with the raw data available, we create a new ground truth which alleviates some of the discovered issues. We also provide recommendations for future data collection. Using the new annotations and ground truth, we benchmark a representative set of bottom-up models for gaze prediction (i.e. those that do not represent the task explicitly). We conclude that while corrected ground truth boosts performance, the implicit representation is not sufficient to capture the effects of task and context on where drivers look. Lastly, motivated by these findings, we propose a task- and context-aware model for drivers' gaze prediction with explicit representation of the drivers' actions and context. The first version of the model, SCOUT, improves state-of-the-art performance by over 80% overall and 30% on the most challenging scenarios. We then propose SCOUT+, which relies on the more readily available route and map information similar to what the driver might see on the in-car navigation screen. SCOUT+ achieves comparable results as the version that uses more precise numeric and text labels.
Open Access
Modular Photovoltaic (PV) Step-up Converter With A Coupled Power Balancing Scheme for DC-Distributed System
(2024-07-18) Kanathipan, Kajanan; Lam, John
The total global energy capacity for renewable energy systems has been increasing exponentially, with photovoltaic (PV) energy having 25% growth rates along with a continual decrease in cost. A high-power PV energy conversion system typically consists of a medium voltage (MV) grid that collects power from individual PV arrays. In a MV Direct Current (DC) distributed grid architecture, since the output voltage of the PV array is significantly lower than the voltage level of medium voltage grid (such as tens of kV), a power electronic interface with sufficiently high voltage gain is required. To safely and effectively connect multiple PV energy sources to the MVDC grid, modular structure of PV power converter is used to convert and maximize the capture of PV energy. The converter consists of external power balancing units to ensure equal power distribution and safe operation amongst all the converter modules. Developing highly power-efficient and cost-efficient power converter topologies and controllers with minimal number of components is the key to achieve a truly optimized PV energy power conversion system. In this dissertation, a highly power-efficient modular PV power converter with high voltage gain and coupled power balancing stages is developed. The first part of this dissertation focuses on the development of a novel current-sensor-less maximum power point tracking (MPPT) technique utilizing a single voltage sensor for the devised high voltage gain PV converter module. In the second part of this dissertation, a new embedded power balancing scheme that utilizes high frequency (HF) interlinking active voltage quadruplers (AVQ) is proposed for the developed modular PV power architecture. The proposed design allows the devised MPPT stage in the PV converter to ensure optimal PV power extraction under all conditions while the interlinking AVQs distribute power equally across all modules to ensure safe operation. In the final part of this dissertation, a power efficiency optimization control scheme is proposed to allow the devised modular PV converter system to achieve high efficiency over a wide range of PV irradiation level. The feasibility of the devised modular PV converter and control concepts are validated through simulation and hardware experiments on proof-of-concept prototypes.
Open Access
Closed-Loop Highly-Scalable Retinal Implant with Fully-Analog ED-Based Adaptive-Threshold Spike Detection and Poisson-Coded Temporally-Distributed Optogenetic Stimulation
(2024-07-18) Yousefi, Tayebeh; Kassiri, Hossein
Intraocular stimulators show promise for treating retinal degeneration by restoring visual input to the damaged retina. This is achieved by capturing images with a wearable camera and accordingly stimulating remaining retinal cells, effectively bypassing dysfunctional photoreceptors. State-of-the-art retinal stimulators face a major challenge due to the lack of cell-type specificity of electrical stimulation (activating both ON and OFF pathways in the retina) leading to limited visual perception due to sending contradictory messages to the brain. This fundamental limit motivated us to investigate the development of an optogenetic-based retinal prosthesis, that uses promoter opsins for selective activation of ON bipolar cells, offering a more natural vision restoration. In developing such a device, the first challenge we faced was optimizing stimulation strategy for optimal therapeutic efficacy. Responding to this challenge, we first present a retina-inspired computational framework to evaluate and optimize an optogenetic epi-retinal neurostimulator. This framework reveals that optical stimulation, compared to electrical stimulation, provides superior visual perception, which improves with increased μLED array resolution. The framework also explores optical stimulation factors and μLED specifications like light intensity and wavelength spatial resolution and light divergence. A critical issue in optogenetics is controlling opsin distribution, as uneven distribution affects light sensitivity across the retina. Variations in tissue properties and fluid dynamics introduce unpredictability in stimulation effectiveness. Our solution to this issue includes a scalable optogenetic stimulator IC, which features channel-specific closed-loop calibration for defining the optimal stimulation intensity using a temporally adaptive-threshold spike detection circuit. The second challenge we addressed was scalability, and by association, energy efficiency of the device. Scaling implantable stimulators is limited by instantaneous power demands during multi-channel stimulation. We address this by exploiting opsins’ sensitivity to integrated optical energy, using Poisson coding, temporally-distributed stimulation to evenly distribute the stimulation power consumption, enabled by our raster scanning technique for efficient μLED addressing. This reduces wireless data communication requirements, and significantly reduces IC-to-optrode interconnections, making large-scale implementation feasible. Our wireless and battery-less stimulator implant comprises blocks for optical stimulation, fully-adaptive spike detection, and closed-loop calibration. It calibrates light intensity for each μLED row based on recorded spiking activity.
Open Access
Modular High-DR Artifact-Resilient Wearable EEG Headset with Distributed Pulse-Based Feature Extraction and Multiplier-Less Neuromorphic Boosted Seizure Detection
(2024-07-18) Dabbaghian, Alireza; Kassiri, Hossein
Wearable electroencephalography (EEG) headsets have emerged as a promising technology with the potential to revolutionize outpatient diagnostics. These devices offer real-time insights into the neurological activity of the brain, paving the way for more precise and personalized treatment strategies. One of the critical factors in the success of these diagnostic headsets is their energy-efficient design, especially for long-term recording applications. Typically, diagnostic headsets consist of multiple active electrodes (AEs) equipped with embedded electronics for functions like amplification and quantization. These electrodes are connected to a central back-end (BE) unit responsible for data processing and, if needed, wireless transmission. A comprehensive analysis of the current state of the art in this field reveals that power consumption in systems with a sufficiently high dynamic range (DR) analog front-end (AFE) and a data-driven classifier, such as a nonlinear support vector machine (NL-SVM) for seizure detection, is mainly dominated by three main components: the AFE (48%), AE-to-BE data communication (26%), and the signal processing required for seizure detection (21%). This underscores the necessity of taking a holistic approach to enhance the efficiency of these key components for an overall energy-efficient design. This study presents design, development and experimental characterization of a channel count- scalable modular fully-flexible platform which achieves motion-artifact-resilient 80 dBDR EEG recording and highly-accurate, energy-efficient seizure detection. The innovation lies in a series of strategies that collectively result in a remarkable 72% reduction in overall power consumption: 1) performing feature extraction (FE) in each AE (i.e., distributed) resulting in 99% reduction in AE-to-BE data communication power, 2) using an ultra-low-power pulse-based method for FE resulting in 92% reduction in its power, 3) performing classification using a neuromorphic multiplier-less boosted-SVM that achieves a comparable accuracy to a resource-intensive NL-SVM while consuming 98% less power, 4) adopting an auto-ranging scheme in the AFE that leads to >17dB DR improvement, 5) employing an AE-specific-boosted distributed feed-forward common-mode rejection (CMR) enhancement method achieving >80dB CMRR in the presence of 1MΩ AE-to-AE input impedance (Zin) mismatch (89dB without mismatch), and 6) an AE-specific electrode-tissue impedance (ETI)-correlation-based circuit for motion artifact detection and suppression.
Open Access
Vulnerability Assessment of Power Transformers and Power Systems to Geomagnetic Disturbances
(2024-03-16) Ariannik, Mohamadreza; Rezaei Zare, Afshin
Powerful solar storms emit plasma that may travel towards the earth. Interactions between the plasma and the earth magnetic field cause geomagnetic disturbances (GMDs), which in turn induce quasi-dc voltage along long conductors in power systems. Assessing the power system resiliency against GMDs requires accurately calculating the induced electric fields and the resultant geomagnetically induced currents (GICs). Offline and online wide-area geomagnetic field monitoring systems are established in this research to estimate GIC flows in power systems accurately. The proposed monitoring systems process the magnetic field signals that are measured at several observatories worldwide. In the offline monitoring system, the magnetic field signals are denoised, and spikes are detected and replaced. The time derivative of the signal is taken by a continuous wavelet transform to prevent amplification of the noises. GICs in a modified IEEE 118-bus benchmark power system are calculated concerning a realistic geomagnetic storm to demonstrate the effectiveness of the proposed signal processing methods. A sliding window is applied in the online monitoring system, and its size is optimized to lower processing time while increasing the signal-to-noise ratio. High amplitude GICs can cause a sharp increase in the hottest-spot temperature of the power transformers. The high temperature allows the formation of gaseous bubbles in the oil-paper insulation and endangers the integrity of the transformer's insulation system. The bubbles include mainly water vapor and emerge in the cavities on the surface of the paper insulation. In the experimental phase of this research, a test setup is created to detect bubbling inception temperature (BIT) for Kraft and thermally upgraded papers (TUPs). The paper samples are dried, prepared at six different moisture levels, and immersed in synthetic ester oil for the experiments. The paper strips are wound around a cartridge heater, and a controller unit raises its temperature at 3 ºC/min and 20 ºC/min rates to detect BIT. The BITs are considered the operational limit on the hottest-spot temperature of the transformers.
Open Access
Neural Spike Compression through Salient Sample Extraction and Curve Fitting Dedicated to High-Density Brain Implants
(2024-03-16) Nekoui Shahraki, Mahdi; Sodagar, Amir
This work proposes a data reduction framework, specific to the compression of extra-cellular neuronal action potentials on brain-implantable microsystems. The proposed framework significantly reduces the extent of data representing spike waveforms, paving the way for the implementation of next-generation, high-density neural recording brain implants. This highly-compressive approach picks a small number of salient samples of the spike, using which and based on some predefined functions the entire spike waveshape is formulated. The amplitudes and timings of the salient samples are sent off the implant in order to reconstruct the spike waveshape on the external side of the system. In addition to exhibiting extremely high data compression capability, this technique is highly hardware efficient, hence it well suits for brain-implantable neural recording microsystems with high channel counts. Based on the proposed framework, a 128-channel neural signal compressor is designed and microfabricated using the TSMC 130-nm CMOS technology, and measures 1.05 mm by 0.35 mm, giving an area-per-channel of 0.00287 mm2. The circuit is tested using a library of intra-cortically recorded neural signals. At an average spike firing rate of 8 Spike/s, the circuit temporally reduces neural data with an average compression rate of ~272, which is equivalent to a true compression rate of ~2176. Operated using a 1-V power supply and at a clock rate of 32 MHz, the 128-channel neural data compressor consumes 0.164 µW/channel.
Open Access
Federated Learning for Heterogeneous Networks: Algorithmic and System Design
(2024-03-16) Wu, Hongda; Wang, Ping
Building reliable machine learning models depends on access to data samples. With the increasingly advanced sensing and computing capabilities on edge devices, the ever-stringent data privacy legislation, and growing user privacy concerns, it is crucial to build learning models from separate, heterogeneous data sources without violating user privacy. Federated Learning (FL) can facilitate collaborative machine learning without accessing user-sensitive data and has emerged as an attractive paradigm for mobile edge networks. However, federated optimization builds on a heterogeneous environment, which brings challenges beyond traditional distributed learning. Though FL is viewed as a promising technique for enabling intelligent applications, the current FL system suffers from high communication costs, restricting it from being applied in mobile edge networks. To fully release the potential, the FL design must be communication-efficient, adaptive, and robust to the heterogeneous training environment. In this thesis, we aim to address the practical challenges of FL in a conscientious manner. Particularly, we try to understand and address some of those challenges in federated networks and build FL systems that fulfill the accuracy, efficiency, and robustness requirements. Starting with the primary challenge, i.e., data heterogeneity, we study how it impacts the model accuracy and communication cost in the collaborative training system. To address this concern, we develop new and scalable algorithms that can quantify the contribution from participating devices, thus alleviating the negative impact of data heterogeneity and reducing the overall communication burden. To handle another major challenge, i.e., the heterogeneity of computation capabilities among different types of edge devices, we devise a new sub-model training method to enable devices with heterogeneous computation capabilities to participate in and contribute to the FL system, making it robust to the straggler effect. The proposed solutions are rigorously compared with popularly adopted benchmarks from theoretical and empirical perspectives. Finally, we provide a preliminary discussion on personalized FL and point out the potentially interesting research directions in the related fields. Although the proposed methods and designs originate from the practical application of FL, the theoretical insights gained from this thesis can be extended to a broader context of trustworthy machine learning.
Open Access
Machine Learning-Based Defences Against Advanced 'Session-Replay' Web Bots
(2024-03-16) Sadeghpour, Shadi; Vlajic, Natalija
The widespread adoption of the Internet has brought about significant benefits for modern society, but has also led to an increase in malicious activities, particularly through the use of web bots. While some bots serve useful purposes, the proliferation of malicious web bots poses a significant threat to Internet security, impacting individuals, businesses, governments, and society as a whole. The emergence of AI-powered web bots capable of mimicking human behavior and evading detection has further exacerbated this problem. This dissertation aims to deepen our understanding of advanced web bots and the web bot attacks that often signal fraudulent online activities. In particular, we focus on session-replay web bots, the latest and most advanced type of web bots, which present an especially difficult challenge in online domains where multiple genuine human users frequently exhibit similar behavioral patterns, such as news, banking, or gaming sites. To achieve our research objectives, we have meticulously curated an extensive dataset encompassing both human and bot-generated data. Additionally, we have developed our own prototype of advanced session-replay bot (the so-called ReBot), which has enabled us to accurately simulate the attacks conducted by this particular category of web bots. Moreover, by infusing randomness into the design of ReBot, we have been able to achieve varying degrees of bot and attack evasiveness. From the defenders perspective, and by leveraging state-of-the-art deep learning algorithms, we have proposed several effective strategies for detection of advanced session-replay bot attacks. One of our proposed techniques deploys the concept of moving-target defence in the form of webpage randomization which is particularly challenging for the attacker to overcome. This thesis also explores the utilization of generative machine learning models for the purpose of generating synthetic bots sessions. The ability to synthesize advance session-replay bots - as opposed to looking for real-world instances of these bots or evidence of their activity in real-world logs - is of critical importance if we are to make timely and effective advances in the field of web bot detection and defence.
Open Access
Integrated Circuits and Systems for Adaptive Optimization of Energy Storage Efficiency in Resonant Inductive Wireless Power Receivers
(2023-12-08) Taghadosi, Mansour; Kassiri, Hossein
Recent developments in highly-miniaturized implantable neuro-stimulators has led to a rapid rise in their required power and data transmission throughput resulting in an increase of instantaneous-to-average ratio in their power consumption. Motivated by crucial role of efficient energy storage in such systems, we introduce energy management strategies in wireless powering links, in which the key performance measure is the energy stored during a limited time interval rather than the average energy delivered to the load. First, the development of an algorithmic scheme for maximizing energy storage in current-mode (CM) resonant inductive power receivers is presented. The efficacy and precision of the presented analytical model is confirmed with CAD-based simulation results and validated using experimental measurements. Furthermore, a 0.45mm2 integrated circuit (IC) fabricated in 0.18µm CMOS is presented that performs the above-mentioned optimization. By continuous monitoring of incident waveform dynamics, the IC automatically adapts its optimal solution on-the-fly to any change in the inductive link's physical or electrical parameters. The computations are implemented using analog circuits which minimize IC's power consumption while making it needless of high-speed ADC/DAC. Our measurement results show that by using the IC, the energy storage efficiency is improved by 53% and 67% for the two tested links, compared to the conservative schemes, while consuming two orders of magnitude smaller energy than it saves through optimization. To the best of our knowledge, this is the first reported link-adaptive calibration-free IC for optimizing energy storage efficiency in CM receivers. Second, a 2×2mm2 IC is fabricated in 0.18µm CMOS that maximizes the energy storage efficiency in resonant inductive links with voltage-mode receivers. The IC automatically stores the maximum possible energy while simultaneously providing the required continuous load's power. In the proposed scheme, the optimal operation is maintained by detecting and operating the receiver at a specific optimal voltage, eliminating the need for direct power measurements and adaptive matching circuits. The power reception and delivery phases are isolated which ensures maximum power reception independent of the actual loading at the receiver. The measurement results demonstrate up to 48.44% and 93.97% improvements for the charging time and the stored power, respectively.
Open Access
Machine Learning and Digital Histopathology Analysis for Tissue Characterization and Treatment Response Prediction in Breast Cancer
(2023-12-08) Saednia, Khadijeh Shirin; Sadeghi-Naini, Ali
Breast cancer is the most common type of diagnosed cancer and the leading cause of cancer-related death in women. Early diagnosis and prognosis in breast cancer patients can permit more therapeutic options and possibly improve their survival and quality of life. The gold standard approach for breast cancer diagnosis and characterization is histopathology assessment on biopsy specimens, which is time and resource-demanding. In this dissertation project, state-of-the-art machine learning (ML) methods have been developed and investigated for breast tissue characterization, nuclei segmentation, and chemotherapy response prediction in breast cancer patients using pre-treatment digitized histopathology images. First, a novel multi-scale attention-guided deep learning model is introduced to characterize breast tissue on digital pathology images according to four histological types. Evaluation results on the test set show the effectiveness of the proposed approach in accurate histopathology image classification with an accuracy of 97.5%. In the next step, a cascaded deep-learning-based model is proposed to delineate tumor nuclei in digital pathology images accurately, which is an essential step for extracting hand-crafted quantitative features for analysis with conventional ML models. The proposed model could achieve an F1 score of 0.83 on an independent test set. At the end, two novel ML frameworks are introduced and investigated for chemotherapy response prediction. In the first approach, a digital histopathology image analysis framework has been developed to extract various subsets of quantitative features from the segmented digitized slides for conventional ML model development. Several ML experiments have been conducted with different feature sets to develop prediction models of therapy response using a gradient boosting machine with decision trees. The proposed model with the optimal feature set could achieve an accuracy of 84%, sensitivity of 85% and specificity of 82% on an independent test set. The second approach introduces a hierarchical self-attention-guided deep learning framework to predict breast cancer response to chemotherapy using digital histopathology images of pre‑treatment tumor biopsies. The whole slide images (WSIs) are processed automatically through the proposed hierarchical framework consisting of patch-level and tumor-level processing modules followed by a patient-level response prediction component. A combination of convolutional and transformer modules is utilized at each processing level. The proposed framework could outperform the conventional ML models with a test accuracy, sensitivity, and specificity of 86%, 87%, and 83%, respectively. The proposed methods and the reported results in this dissertation are steps toward streamlining the histopathology workflow and implementing response-guided precision oncology for breast cancer patients.
Open Access
State Estimation of Li-ion Batteries Using Machine Learning Algorithms
(2023-03-28) Babaeiyazdi, Iman; Rezaei Zare, Afshin
Lithium-ion batteries are mainly utilized in electric vehicles, electric ships, etc. due to their virtue of high energy density, low self-discharge, and low costs. Electric vehicles are prone to accelerated battery degradation due to the high charging/discharging cycles and high peak power demand. Hence, efficient management of the batteries is a dire need in this regard. Battery management systems (BMS) have been developing to control, monitor, and measure the variables of the battery such as voltage, current, and temperature, to estimate the states of charge (SOC) and state of health (SOH) of the battery. This study is divided into three parts; in the first part, the SOC of the battery is estimated utilizing electrochemical impedance spectroscopy (EIS) measurements. The EIS measurements are obtained at different SOC and temperature levels. The highly correlated measurements with the SOC are then extracted to be used as input features. Gaussian process regression (GPR) and linear regression (LR) are employed to estimate the SOC of the battery. In the second part of this study, the EIS measurements at different SOC and temperature levels are employed to estimate the SOH of the battery. In this part, transfer learning (TL) along with deep neural network (DNN) is adopted to estimate the SOH of the battery at another outrange temperature level. The effect of the number of fixed layers is also investigated to compare the performance of various DNN models. The results indicate that the DNN with no fixed layer outclasses the other DNN model with one or more fixed layers. In the third part of this dissertation, the co-estimation of SOC and SOH is conducted as SOC and SOH are intertwined characteristics of the battery, and a change in one affects the other variation. First, the SOH of the battery is estimated using EIS measurements by GPR and DNN. The estimated SOH, along with online-measurable variables of the battery, i.e., voltage and current, are then utilized as input features for long-short term memory (LSTM) and DNN algorithms to estimate the SOC of the battery.
Open Access
Machine Learning and Quantitative Imaging for the Management of Brain Metastasis
(2023-03-28) Jalalifar, Seyed Ali; Sadeghi-Naini, Ali
Significantly affecting patients’ clinical course and quality of life, a growing number of cancer cases are diagnosed with brain metastasis annually. Although a considerable percentage of cancer patients survive for several years if the disease is discovered at an early stage while it is still localized, when the tumour is metastasized to the brain, the median survival decreases considerably. Early detection followed by precise and effective treatment of brain metastasis may lead to improved patient survival and quality of life. A main challenge to prescribe an effective treatment regimen is the variability of tumour response to treatments, e.g., radiotherapy as a main treatment option for brain metastasis, despite similar cancer therapy, due to many patient-related factors. Stratifying patients based on their predicted response and consequently assessing their response to therapy are challenging yet crucial tasks. While risk assessment models with standard clinical attributes have been proposed for patient stratification, the imaging data acquired for these patients as a part of the standard-of-care are not computationally analyzed or directly incorporated in these models. Further, therapy response monitoring and assessment is a cumbersome task for patients with brain metastasis that requires longitudinal tumour delineation on MRI volumes before and at multiple follow-up sessions after treatment. This is aggravated by the time-sensitive nature of the disease. In an effort to address these challenges, a number of machine learning frameworks and computational techniques in areas of automatic tumour segmentation, radiotherapy outcome assessment, and therapy outcome prediction have been introduced and investigated in this dissertation. Powered by advanced machine learning algorithms, a complex attention-guided segmentation framework is introduced and investigated for segmenting brain tumours on serial MRI. The experimental results demonstrate that the proposed framework can achieve a dice score of 91.5% and 84.1% to 87.4% on the baseline and follow-up scans, respectively. This framework is then applied in a proposed system that follows standard clinical criteria based on changes in tumour size at post-treatment to assess tumour response to radiotherapy automatically. The system demonstrates a very good agreement with expert clinicians in detecting local response, with an accuracy of over 90%. Next, innovative machine-learning-based solutions are proposed and investigated for radiotherapy outcome prediction before or early after therapy, using MRI radiomic models and novel deep learning architectures that analyze treatment-planning MRI with and without standard clinical attributes. The developed models demonstrate an accuracy of up to 82.5% in predicting radiotherapy outcome before the treatment initiation. The ground-breaking machine learning platforms presented in this dissertation along with the promising results obtained in the conducted experiments are steps forward towards realizing important decision support tools for oncologists and radiologists and, can eventually, pave the way towards the personalized therapeutics paradigm for cancer patients
Open Access
Transformer Thermal Assessment under Geomagnetically Induced Current Conditions
(2023-03-28) Akbari, Milad; Rezaei Zare, Afshin
Power transformers are one of the most critical and expensive pieces of equipment in power systems. The widespread use of the transformer in power grids and its high cost make the lifetime and reliability of this apparatus highly important. Although some factors affect the reliability of transformers, thermal and electrical stresses are the main reasons for transformer failure. As a result, transformer thermal modeling is essential in the design and operation stage to represent the thermal behavior of transformers during normal operation or transient phenomena. However, the multi-physics behavior of transformers and the nonlinear and frequency-dependent parameters make this modeling a challenging task. This thesis aims to develop a more accurate transformer model for representing the thermal behavior of transformers, especially during transient phenomena such as Geomagnetically Induced Current (GIC). To fulfill this goal, it is necessary to perform several tasks in different fields, such as geometry and material modeling, electromagnetic studies, and investigation into computational fluid dynamic (CFD) analysis of transformers. First, the GIC phenomenon and its impact on the transformer are briefly explained. Then, a comprehensive literature review of existing transformer thermal models is performed to find their drawbacks. A 3-phase, 3-leg transformer is then subjected to an electromagnetic-thermal study in both normal and GIC conditions. It is shown that the structural parts, including the tank, clamps, and tank shunts, are saturated with a small amount of GIC. However, the transformer core becomes saturated with larger currents, resulting in additional stray losses in the structural parts. The findings show that the most vulnerable part is the tank, as the hot spot temperature (HST) of the tank approaches 372.14 0C, which is double the permissible limit, under 66.6A GIC per phase. Finally, a new approach is proposed to determine the HST of OIP bushing based on the FEM-modified thermal equivalent circuit (TEC) model. The proposed model can accurately estimate the HST of the bushings under normal and GIC conditions. Furthermore, a detailed thermal analysis is performed to investigate the impact of different parameters such as load, ambient temperature, and top oil temperature on the thermal performance of bushings.
Open Access
Novel Design and Energy Management Approaches for Seamless Integration and Adoption of Plug-In Electric Vehicles
(2022-12-14) Al-Obaidi, Abdullah Azhar Abdullah; Farag, Hany E. Z.
Electric vehicles (EVs) are witnessing increased utilization throughout the world as an alternative to fossil-fueled vehicles. However, the adoption of EVs and their integration into the power grid is yet to be fully materialized due to several issues, of which two are the most salient. First, the extensive deployment of EVs can bring challenges to the grid if not properly managed. Second, access to a variety of EV supply equipment (EVSE) in different areas is still lacking. To that end, the research in this thesis aims to address these issues through the development of adaptive approaches that enhance the management of EV energy and the development of a charging strategy and a design approach that help to expand the proliferation of EV charging infrastructure. Three approaches that are adaptive to their operator/user preferences are developed to enhance energy management in EVs. The first approach allows adaptive utilization of EV batteries' distributed energy resources in an EV fleet system for concurrent services to the transportation sector and ancillary services market. The second approach is a decentralized quality of service (QoS)-based scheme for peer-peer (P2P) energy trading among EV energy providers and consumers. The proposed mechanism is designed to match energy traders based on consumers' and providers' QoS requirements and offers, respectively. The third approach is a bidirectional smart charging algorithm for EVs considering P2P energy trade, provision of ancillary services to the grid, and utilization of low electricity prices for battery charging. The algorithm incorporates user preferences into the scheduling process enabling it to adapt to various conditions. Further, to expand the proliferation of EV charging infrastructure, this thesis introduces (i) a charging control strategy that does not require a communication network, which in turn reduces additional grid upgrades, and (ii) a design approach for EV parking lots that helps private investors to participate in the growth of charging facilities. The findings of this thesis highlight the efficacy of the proposed approaches in achieving their objectives. This would provide implementable and cost-effective solutions to facilitate EVs deployment and address imminent and timely concerns that limit the wide adoption of EVs into electric distribution infrastructure.