M.S. Theses

Permanent URI for this collection

https://digitalarchive.library.bogazici.edu.tr/handle/123456789/11647

Browse

Now showing 1 - 20 of 471

Age of information in network coded internet of things
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2023., 2023) Kahraman, İbrahim.; Koca, Mutlu.
The unique nature of the Internet of things (IoT), with potentially millions of interconnected devices featuring varying data rates, power levels, bandwidth, and range specifications, demands different performance metrics compared to traditional communication systems. In conventional wireless communication like cellular networks, performance markers such as data rate and spectral efficiency are paramount. However, in energy-constrained real-time IoT applications with low data rates, the timeliness of information, measured as age of information (AoI), takes center stage. AoI represents the time elapsed since the last packet update originated at its source and has garnered significant research attention. In this regard, This thesis provides an overview of AoI’s role in designing and optimizing IoT applications, including AoI-based optimization, scheduling for IoT networks, application of learning methods in large-scale IoT systems, real-life applications and experimental results, together with a synopsis of potential future applications and research challenges, is provided in this thesis. Additionally, the timeliness in delivering updates within a multi-source multi-hop IoT networks via multicast transmissions with or without employing network coding is considered in this thesis. The effect of network coding on the average AoI is investigated employing a completely probabilistic model in a two-stage transmission scheme. The theoretical findings demonstrate that network coding has great potential to improve data freshness in multi- source multi-hop IoT networks, which closely represent the spine of real-life scenarios, and extensive simulation results corroborate the theoretical findings.
Inter-numerology interference minimization ın 5G : a deep reinforcement learning based approach
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2023., 2023) Erk, Tuğrul Can.; Pusane, Ali Emre.
The use of 5G technologies has become prevalent over the years with the increased use of mobile devices, video services, and more. At the same time, 5G technology brings lower latency, higher reliability, and higher throughput than 4G technology. creation of network slices via software-defined networks and network virtualization functions enables these improvements. A network slice provides flexibility to the network. Each slice can be dynamically configured within itself via SDN/NVF. The network meets diverse requirements of diverse services by creating a network slice. At the higher level, NVF is responsible for creating and managing network slices. The reflection of a network slice on the physical layer is RAN slicing. With the numerology concept merging with 5G, the radio spectrum and corresponding resources become configurative in a manner of bandwidth, sub-carrier spacing, etc. However, the numerology solution has one drawback. Changing sub-carrier spacing in the resource grid destroys the orthogonality principle in the traditional OFDM signals. This leads the network to face new interference, inter-numerology interference. The solutions that enable these improvements have also brought optimization problems that we have not faced In 4G networks. The main problem is about optimization since 5G technology requires heterogeneous signals for heterogeneous services. Therefore, the optimal allocation of limited resources in order to prevent this problem is the essential aim of this paper, with the minimization of inter-numerology interference and maximizing channel capacity.
Revisiting image captioning structures based on CNN and RNN, and improving the performance using modified decoders with residual connections
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2023., 2023) Saraçoğlu, Sinan.; Anarım, Emin.
In this thesis, the image captioning structure consisting of a Convolutional Neural Network (CNN) as the encoder and a Recurrent Neural Network (RNN) as the decoder is visited by comparing and evaluating the effects of different image feature extractors, different RNN cells, different types of word embeddings, and the involvement of residual connections between the RNN cells. The famous “Show, Attend and Tell” model is modified by adding residual connections between the RNN cells and adding other modifications on both the encoder and the decoder side, which improved the performance of the model on the image captioning task. Furthermore, models were trained by implementing 3 different pre-trained word embeddings and their benefits were explored. With the best model, 34 BLEU-4 points and 15 SPICE points improvement were achieved compared with the base model. The effects of training our best model with the images transformed into the frequency domain rather than the images represented in the spatial domain are investigated and it is concluded that this approach cannot enhance the performance of the model. The results of the experiments demonstrate the effectiveness of the proposed modifications and provide insights into the potential of residual connections.
A novel scheduling strategy for priority-aware iot networks for age of information optimization
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2023., 2023) Sayınbaş, Oğuzhan.; Anarım, Emin.
As the need for wireless communication grows, the need to increase spectral efficiency and reduce latency in communication has become more critical. In this study a novel scheduling strategy is proposed based on solving the Knapsack problem in Internet of Thing networks. The main goal is to minimize the difference between Age of Information (AoI) values of sensors and Peak AoI (PAoI) constraints subject to average and peak transmission power constraints given the communication resources are scarce. At first, we formulated the problem according to the specified PAoI constraint, transmission power, and frequency band constraints. The proposed approach involves reformulating the original problem as a Knapsack problem. This is done by assigning a value as the decrease in AoI of a sensor if a status update transmission is successful. In addition, a weight is specified as the frequency band amount that a sensor requires to transmit its update. Additionally, Fully Polynomial-Time Approximation scheme (FPTAS) is proposed to decrease the computational complexity while preserving the quality of results. Secondly, algorithms of benchmark methods were created together with the proposed method. Finally, the results of the simulations completed using the adapted algorithms are given. The proposed scheduling method is shown to outperform the benchmarks which are Multi-Armed-Bandit Q- learning and Whittle’s Index strategies.
Detection and mitigation techniques against eavesdropping and spoofing in GPS
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2023., 2023) Demir, Mehmet Özgün.; Pusane, Ali Emre.; Karabulut Kurt, Güneş
As a significant CPS application, V2X networks have their own specific security atmosphere in addition to their latency and energy efficiency requirements. There are two critical attack types against secure V2X deployments, which are eavesdropping and spoofing attacks. In this thesis, we evaluate these attacks in a novel system model, where aerial spoofers, ground-located spoofers, and eavesdroppers are jointly integrated. Against eavesdropper attacks, an error vector is added to the information data to falsify the attackers using FEC codes with respect to and relay-aided McEliece cryptosystem-based methodology. The results show that significantly noisier channels between relays and the eavesdropper enhance information security. Additionally, the impacts of time spoofing attacks on satellite positioning signals are studied. For this purpose, two pseudorange-based spoofing detection algorithms with different performances and complexity levels are proposed to mitigate ground-located spoofers. In order to find the best detection thresholds of these algorithms, Pareto fronts are plotted, and then a decision tree-based detection approach is analyzed. These algorithms are also tested against novel aerial spoofers, which suffer additional mobility and atmospheric errors. Finally, a high-level hybrid decision methodology is applied to improve spoofing detection rates and reduce false alarm rates by integrating a decision fusion module. With this approach, the decision fusion module combines individual outcomes of the proposed algorithm and improves spoofing decision performance. In summary, this thesis proposes extensive strategies against eavesdropper and spoofing attacks in V2X networks by providing secure transmission schemes and detection algorithms in a novel system model.
Architectural exploration of FPGAS and RTL2GDSII implementation of an FPGA
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2023., 2023) Eroğlu, Mehmet Sait.; Başkaya, Faik.
Growing design complexity and cost has forced designers to build programmability into System-on-Chip (SoC) designs to reduce the number of costly chip re-spins and amortize IC costs over several fabrications. Programmability of embedded FPGA (eFPGA) cores is one of a handful of design solutions to meet this challenge. However, creating a new FPGA is challenging because of the significant effort that must be spent on circuit design, layout, and verification. The time required until the tape-out is approximately 1 year of a large team from architecture definition to tape-out for a new FPGA, since the process is primarily done manually. Researchers have developed automated methodologies to overcome these barriers by modeling FPGA fabrics as Verilog netlists and generating layouts using ASIC automated design tools. Simplifying and shortening the design process would be advantageous since it could reduce the time to implement eFPGAs in SoCs while enhancing architecture explorations. For this purpose, OpenFPGA is introduced, an open-source framework that enables automated prototyping for FPGA architectures. To enable various design purposes, OpenFPGA integrates several popular open-source EDA tools, i.e., VTR and Yosys, with its own custom tools. In this work, we designed an FPGA using OpenFPGA framework and investigated the issues faced in the architecture selection, circuit design, layout, and verification of such a FPGA with T¨urkiye’s domestic 250nm CMOS technology.
Deep learning based automatic modulation classification for sub- carriers of OFMD signals
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2023., 2023) Tosun, Gökhan.; Pusane, Ali Emre.
Automatic modulation classification (AMC) is automatically identifying and classifying the modulation schemes employed in digital communications. By accurately identifying the modulation scheme, AMC enables communication systems to adapt their parameters, optimizing efficiency, spectral utilization, and overall performance. The majority of the literature on AMC focuses on the single-carrier communications systems. This thesis addresses the gap between the AMC and multi-carrier communications systems. Two architectures are proposed. Both employ a filter bankconvolutional neural network (CNN) complex. The first architecture uses raw features and a maximum operation to perform classification, whereas the second architecture learns feature patterns by employing a fully connected neural network (FNN). It is observed that the raw features are not sufficiently informative for theoretical and practical purposes. It is further observed that putting together the raw features and allowing the transformations on the combinations of the raw features, effectively forming a decision context, improves the performance significantly. The performances of both architectures are analyzed through the accuracy metric and confusion matrices. Finally, the thesis is concluded by summarizing the experiments, results, and implications and mentioning the possible future work.
Hardware implementation of montgomery multiplier based low-power fips-compliant random prime number generator
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2023., 2023) Kaysici, Halil İbrahim.; Başkaya, Faik.
The Internet of Things (IoT) connects devices, vehicles, buildings, and other objects to the internet, improving efficiency and automation in various applications. More connected devices require secure communication and data storage to protect data privacy and integrity. Public-key cryptography, which uses two mathematically related keys to encrypt and decrypt data, is better for IoTs and is being used more. Secure hardware and ASIC design create tamper-resistant, energy- efficient devices resistant to physical and logical attacks. Secure IoT communication requires open-source cryptography algorithms and hardware accelerators. This thesis explains the design of an IoT SoC to generate random prime numbers efficiently. We design secure and compatible hardware using standard specifications and test suites. Parametrized module design allows flexible and scalable hardware design. Pipelining and source-sharing configurations allow us to observe area- latencybandwidth tradeoffs. Since IoT includes a wide range of hardware, observing different configurations is useful. Designed hardware also interacts with software to benefit from the hardware-software co-design approach. This thesis proposes two new RNG designs and implements a scalable Montgomerybased modular multiplier. The modular multiplier forms a basis for Miller-Rabin and Lucas probabilistic primality tests. The final hardware design combines the proposed RNGs and primality tests with an open-source Ibex core into an IoT SoC.
SDR-based tracking system for biodegradable implant sensor
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2023., 2023) Aydın, Diren Erdem.; Dumanlı Oktar, Sema.
Covid-19 pandemic had an impact on society’s priorities towards the end of 2019. In the short term, people had to relinquish their regular social activities and adopt new habits to safeguard their health. Consequently, personal health emerged as the foremost concern in daily life, surpassing all others. This heightened awareness of personal health has led to an increased demand for novel medical systems, irrespective of whether they are specifically related to the Coronavirus or not. Even prior to the recent pandemic, advancements in nano and RF technologies have already hinted at the potential of health applications based on wireless communication. RF medical imaging and sensing has emerged as a promising adjunct to existing diagnostic tools in the field of medical applications, primarily due to its cost-effectiveness, portability, and non- ionizing radiation properties. Software Defined Radios (SDR) are an exceptionally well-suited option for medical applications due to their remarkable flexibility, enhanced RF specifications, and compact hardware dimensions. This thesis introduces a novel low-cost and wearable system as a viable substitute for the Vector Network Analyzer (VNA), a pivotal measurement technology in the RF industry. While several studies have explored the use of SDR as a substitute for VNAs, it is noteworthy that this research represents the pioneering application of an SDR-based system in conjunction with genetically modified bacteria for sensing biological processes and specific chemical outputs generated by molecular communications. The system design and implementation involve the utilization of GNU Radio and Python for Adalm-Pluto SDR development, MATLAB for advantaged signal processing tasks, and the establishment of a measurement setup using biological phantoms. The comparison of SDR results with those from a VNA provides valuable insights into the system’s effectiveness.
Coverage optimization for 6G using reconfigurable intelligent surfaces
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2023., 2023) Keşir, Samed.; Pusane, Ali Emre.
Recent research has shown that a reconfigurable intelligent surface (RIS), which is an attractive technology due to its low complexity and power requirements, can considerably improve the performance of wireless communication systems. RIS composed of small, reconfigurable, passive elements can be programmed to reflect the signal in a particular direction, amplify or attenuate it, or even introduce additional signal components to improve the overall communication performance. RIS can be utilized for different purposes including enhancements to the coverage, positioning, capacity, security, and sustainability. In this thesis, coverage enhancement and the physical layer security (PLS) use cases of RIS are investigated. For the coverage enhancement, single-user and two-user scenarios are discussed. In the single-user scenario, a RIS optimization method using Convolutional Neural Networks (CNN) is proposed to reduce the number of iterations required for optimization. The results show that the proposed CNN model significantly decreases the number of steps required for the configuration by causing a slight degradation in the performance. In the two- user scenario, both coverage enhancement and security use cases are investigated. For coverage enhancement, it is shown that the RIS gain distribution between the users can be managed with respect to the needs by configuring the RIS accordingly. For the security use case, measurement-based characterization of RIS for providing physical layer security, where the transmitter (Alice), the intended user (Bob), and the eavesdropper (Eve) are deployed in an indoor environment is presented. Also, the measurement results are verified with the simulation results. Finally, a thresholding method is proposed to prevent Eve’s dominance in PLS and provide more RIS gain to Bob in return for a slight reduction in secrecy capacity.
Text-independent speaker verification with very short utterances
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2023., 2023) Ülgen, İsmail Rasim.; Arslan, Levent M.
The accuracy of the text-independent speaker verification suffers greatly when the speech duration is very short. In this thesis, some methods are proposed aiming to compensate for the drastic performance degradation in speaker verification with very short utterances. Firstly, methods that try to leverage the additional information from large-scale speaker datasets are proposed in order to enhance the limited speaker information that is present in the very short speech utterances. Secondly, the problem of short utterances is tackled in a more specific way in terms of the phonetic content of the speech. An analysis of phonetic mismatch between verification utterances is performed, along with experiments of a back-end scoring module that is aware of the phonetic mismatch in speaker verification. Furthermore, contributions to the speaker verification in general, which might be applicable to the very short duration conditions are presented. A novel loss function for back-end scoring module training is introduced. The proposed loss function outperformed the baseline loss function in all cases, including very short duration scenario. Lastly, a novel unsupervised domain adaptation of the discriminative back-end scoring for speaker verification is proposed. The proposed adaptation method improved the performance of the out-of-domain backend scoring model in the target domain in all cases. The relative improvement of the proposed method, compared to baseline adaptation methods, is highest in short duration conditions. NOTE Keywords : speech processing, deep learning, speaker recognition.
Design and simulation of an auscultation data acquisition system on a single integrated chip
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2022., 2023) Kaya, Kerem.; Kahya, Yasemin.; Dündar, Günhan, 1959- .
The COVID-19 pandemic has reminded us of the importance of remote healthcare and early detection of symptoms, as most healthcare systems cannot accommodate all patients at once. The burden on health systems can be reduced with smart wearable remote sensing technologies and online health services. This study aims to design a wearable auscultation data acquisition system on a single integrated chip for remote health monitoring. In this study, available auscultation devices, from a conventional stethoscope to multi-channel computerized data acquisition systems, were carefully investigated. The current standard in computerized auscultation data acquisition systems was analyzed and the design choice for a sound acquisition system was presented. The use of low current consuming, high performance MEMS microphone systems has been listed in other wearables in terms of active noise cancellation and as a potential candidate for the proposed low-power wearable auscultation sensor interface. The versatile switched-capacitor-based gain stage and second-order lter design is quickly implemented with Python support. In order to place the other building blocks of the designed new auscultation data acquisition system in a single system, the post-layout simulation of the switched capacitor lter blocks was made using TSMC 180 nm technology, and an important progress is made in development of the proposed system in a single integrated chip.
An efficient flash translation layer implementation for nor flash memory
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2023., 2023) Çam, Fatih.; Başkaya, Faik.
As technology develops day by day, the amount of data that electronic systems need to manage is increasing in direct proportion. Especially in embedded systems, the necessity to use space efficiently does not allow flash memories used as storage units to exceed a certain size. The fact that the technology on which flash memories are built has a certain lifetime has made it necessary to use these technologies in the most effective way. The algorithm developed as a solution to this problem is the flash translation layer (FTL). Different FTL algorithms have been proposed over the years, and these algorithms have led to some disadvantages along with many advantages. The aim of this thesis is to present a much simpler and implementable FTL algorithm, which is different from the complex algorithms proposed before, and which also shows high success. For read and write operations, the target page address is found directly from the mapping table and the operation is performed. In the write operation, the target page address is determined by the value from the random number generator. If data has been written to this flash address before, the data at this address is moved to the next free flash address and the data to be written is written to the page address from the random number generator. With this method, page writing is carried out quickly and effectively. The algorithm was implemented on the FPGA and tested over UART with test software designed on the computer. The results of the test application showed that the designed algorithm used flash memory pages with a regular distribution and the success rate was 98%. In addition, only one flash memory page sacrifice was sufficient to use this algorithm. Thus, this study did not cause space loss in flash memory, unlike other algorithms.
Design of memory encryption and authentication for secure IoT edge devices
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2023., 2023) Günay, Recep.; Başkaya, Faik.
The security of computer systems has become very important as the Internet of Things (IoT) technology has improved and the number of electronic devices in our daily lives has increased dramatically. üne particular weakness of these devices is the off-chip memory interface since they are easily accessible. They have been subject to various attacks focusing on this weakness such as cold-boot attack and replay attack. Most of the solutions in the literature try to solve this issue by memory encryption and memory authentication with high performance and high hardware cost using cryptography algorithms like AES and SHA. A secure memory solution with memory encryption and authentication with low area and power consumption cost is designed in this thesis. ASCON, a finalist in the NIST lightweight cryptography standardization contest, is used for encryption and hash function. Using a single hardware block for both functions reduces the hardware cost with respect to the literature. A system on chip (SoC) is designed consisting of a secure memory controller with ASCON and metadata cache, key generation block with a built-in true random number generator (TRNG), and secure on-chip storage slots around the open-source RISC-V processor PICORV32. The performance and power costs of encryption and authentication are reduced by applying cache snoops during re-encryption and tree traversal. The SoC is designed in Verilog and implemented in FPGA for hardware verification. It has low area and power consumption overhead with reasonable storage overhead and acceptable performance reduction for IoT applications. NOTE Keywords : Digital circuits, Integrated circuits, Logic circuit, Data encryption.
The known network attack detection and unknown network attack identification based on deep learning methods
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2023., 2023) Ateş, Pelin Damla.; Anarım, Emin.
As accessing information has become easier, the encountered threats along the way have become more frequent. When it was realized that these ways of accessing information were not sufficiently reliable, various prevention techniques were implemented. One of the most useful of these measures is the intrusion detection systems. The Intrusion Detection Systems provide a comprehensive analysis of the network. This way, potential threats can be rapidly detected as quickly as possible and the necessary measures can be taken. On the other hand, classifying network traffic is not only important for identifying threats but it is also crucial for gaining insight into the network’s overall behaviour. Utilizing generative networks for these purposes has become one of the most popular methods in recent times. In this study, deep learning methods are employed to analyze network traffic by using autoencoder models. In the proposed method, network traffic analysis consists of two stages. The first stage aims to correctly classify the classes in the training data. The second stage focuses on detecting unknown classes which is achieved through the application of Extreme Value Theory. Thanks to this mathematical approach, successful separation of known and unknown classes is achieved. The utilized data can be evaluated under two different headings. The first one consists of network attack types while the second comprises popular social media traffics. According to the performance evaluation metrics, the proposed procedure demonstrates satisfactory results in both the classification of known classes and the detection of unknown classes.
Design and radiation analysis of various analog to digital converters
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2023., 2023) Türker, Atakan.; Dündar, Günhan, 1959- .; Ozanoğlu, Kemal.
This study is on the schematic level design of various ADCs and their radiation analysis. In this context, ADCs with three different architectures, which are flash ADC, SAR ADC, and sigma-delta ADC, are designed and tested under temporary and permanent radiation effects. The flash ADC is designed in Verilog-A, which is a modeling language for analog circuits, whereas SAR ADC and sigma-delta modulator are designed by utilizing transistors in 65nm CMOS technology. The designed ADCs are examined under the SET effect, which causes temporary effects on the circuit, and the TID effect, which causes permanent effects. The designed flash ADC is examined only under the effect of TID, and the results of this effect are presented. For the SAR ADC, designs are made using two different capacitive DAC topologies to be able to compare with each other. These SAR ADCs are examined under the effects of TID and SET separately, and the results are explained. As a result of SET simulations of ADCs, the radiation performances of the designed ADCs are compared, and the sensitive nodes of the DACs are determined. In the sigma-delta ADC part, a first- order modulator is designed and investigated under the effects of TID and SET. In addition to the presented results, the sensitive nodes of the integrator are determined, and various inferences are made.
Mode-S radar interrogation algorithm design, simulation environment setup and error correction
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2023., 2023) Aydın, Ahmet Günhan.; Öncü, Ahmet.
The air traffic density was relatively low back in the day compared to today’s dense air traffic environment. This increasing trend in air traffic density will continue in the near future with the addition of different aerial vehicles. Before the Mode-S protocol, Mode A and Mode C were in use; however, the Mode A/C configuration was only usable in sparsely dense air traffic. One of the useful features of Mode-S is the ability of probabilistic interrogation. However, there has not yet been a sophisticated algorithm for closely located large numbers of aircraft. Considering a futuristic air environment with a swarm of drones and airbuses equipped with transponders, the probabilistic interrogation feature of Mode-S is utilized, and an algorithm is designed. The proposed algorithm is able to collect close aircraft information in a relatively short time. There has also been created a high-level Mode-S uplink and downlink communication simulator in order to exchange all-call communication and record the algorithm’s performance in terms of time and number of interrogations sent. Furthermore, error correction of erroneous messages, usually from garbling and noise, is investigated. An error correction method is developed and tested with real-time data coming from commercial aircraft.
3D-engineered muscle tissue as a wireless sensor
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2023., 2023) Karabulut, Çağla.; Dumanlı Oktar, Sema.
Implantable and wearable biomedical devices are advancing with new sensor technologies, holding great potential for early disease detection through continuous, real-time monitoring of physiological parameters. However, the majority of existing biomedical devices have limited lifetimes due to their power requirements and often focus on monitoring physical parameters rather than specific molecules relevant to specific diseases. The work detailed in this thesis proposes a wireless sensing and communication platform that can achieve in-vivo, real-time sensing at a molecular level by utilizing engineered mammalian cells. The proposed platform consists of a cell-based bio- hybrid implant device and a dual-port, wide-band on-body antenna. The molecular sensing is achieved by the bio-hybrid implant that is composed of three main components: a flexible scaffold, an in-body passive implant antenna, and 3D-engineered muscle tissue. The genetic circuitry of the cells that make up the 3D-engineered muscle tissue can be manipulated. This manipulation makes the tissue responsive to specific target molecules and the presence of these molecules triggers a contraction in the tissue. The tissue contraction and relaxation are used to reconfigure the resonance frequency of the implant antenna that is located on the flexible scaffold. To monitor the changes in resonance reconfiguration, the on- body reader antenna is positioned outside of the human body. The implant antenna’s resonance variations are observed in response to the presence of the molecule of interest. In this thesis, the bio-hybrid implant and the on-body reader antenna were designed and fabricated. The sensing system is mechanically and electromagnetically simulated. Based on the simulations, electromagnetic measurements were taken inside tissue-mimicking phantoms to track implant antenna reconfiguration.
A wearable antenna system for structural monitoring of biodegradable magnesium-based suture wires
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2022., 2022) Sezgen, Ozan Furkan.; Dumanlı Oktar, Sema.
Open-heart surgery is a way of treating various heart problems. It is a challenging operation during which the operating surgeon will cut through the sternum and spread the ribcage to access the heart. After sternotomy, complications such as sternal separation and wound or subcutaneous infections may occur due to the sensitivity of the operation region. The probability of occurrence for these post-operative infections is low; however, the mortality rate is high in case of infection. Such risks make remote monitoring and examining patients’ health more critical in the post-operative period. Computed tomography appears as the traditional way of tracking the sutures surrounding the sternum. Stainless steel and titanium are popular for sutures to close the sternum after open-heart surgery. This thesis proposes an alternative suture material and a novel structural real-time monitoring technique in the post-operative period. Magnesium-based sutures have become significant in medical applications due to their biocompatibility and its ability to degrade without releasing any byproduct. A wearable antenna system is a possible solution to monitor patients who underwent open-heart surgery. Ultra- wideband(UWB) coplanar waveguide(CPW) fed disc monopole antenna is chosen as the on-body reader antenna and is optimized to operate on the human body and is fabricated on 1.27 mm Rogers RO3210 with relative permittivity of 10.2. Next, a series of analyses are made in the numerical model, such as the degradation tracking of magnesium-based sutures, breakage points on the suture, different thicknesses to represent overweight and underweight patients, and antenna position effect. After that, the measurement setup is established, including the tissue-mimicking human average and human bone phantoms. Finally, simulation and measurement results are compared and interpreted.
Resilient distributed algorithms for solving linear algebraic equations in faulty networks
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2022., 2022) Çiftçi, Oğuzhan.; Akar, Mehmet.
Various methods have been developed to solve linear algebraic equations distributively over multi-agent networks. Most studies consider that all agents are trustworthy and utilize all the received data from their neighbors throughout the process. Nevertheless, cooperation between non-faulty agents is disrupted if faulty agents intrude into the network. This thesis aims to develop algorithms to detect all faulty agents in the network without prior knowledge of the number of faulty agents. We study four fault models: random-state, fixed-state, single-faced, and double-faced and propose fault detection procedures according to the characteristics of these fault models. First, we introduce a method in which each agent can determine its neighbors’ system of equations if it receives sufficient solution estimations from neighboring agents. By utilizing this method, we propose a synchronous discrete-time distributed detection algorithm for the perfectly synchronized agents in terms of their event times. On the other hand, the event time sequences of different agents are not always assumed to be synchronized. Therefore, we also propose an asynchronous discrete-time distributed fault detection algorithm to analyze the effect of the asynchronous event times of agents. Also, we discuss the applicability of our detection algorithm in continuous-time systems. Moreover, complexity analyses for the proposed algorithms are carried out. Theoretical results are also illustrated by numerical examples.

Browse

Recent Submissions