LEE- Telekomünikasyon Mühendisliği-Doktora
Bu koleksiyon için kalıcı URI
Gözat
Son Başvurular
1 - 5 / 15
-
ÖgeCompressive sensing of cyclostationary propeller noise(Graduate School, 2023-09-12)This dissertation is the combination of three manuscripts -either published in or submitted to journals- on compressive sensing of propeller noise for detection, identification and localization of water crafts. Propeller noise, as a result of rotating blades, is broadband and radiates through water dominating underwater acoustic noise spectrum especially when cavitation develops. Propeller cavitation yields cyclostationary noise which can be modeled by amplitude modulation, i.e., the envelope-carrier product. The envelope consists of the so-called propeller tonals representing propeller characteristics which is used to identify water crafts whereas the carrier is a stationary broadband process. Sampling for propeller noise processing yields large data sizes due to Nyquist rate and multiple sensor deployment. A compressive sensing scheme is proposed for efficient sampling of second-order cyclostationary propeller noise since the spectral correlation function of the amplitude modulation model is sparse as shown in this thesis. A linear relationship between the compressive and Nyquist-rate cyclic modulation spectra is derived to utilize matrix representations for the proposed method. Cyclic modulation coherence is employed to demonstrate the effect of compressive sensing in terms of statistical detection. Recovery and detection performances of sparse approximation algorithms based on greedy pursuits are compared. Results obtained with synthetic and real data show that compression is achievable without lowering the detection performance. Main challenges are weak modulation, low signal-to-noise ratio and nonstationarity of the additive ambient noise, all of which reduce the sparsity level causing degraded recovery and detection performance. Higher-order cyclostationary statistics is introduced to characterize propeller noise due to its non-Gaussian nature. The third-order cyclic cumulant spectrum, also known as the cyclic bispectrum, is derived and its sparsity is demonstrated for the amplitude modulated propeller noise model. Cyclic modulation bispectrum is proposed for feasible approximation of the cyclic bispectrum based solely on the discrete Fourier transform. Additionally, compressive sensing of the cyclic modulation bispectrum is suggested. Numerical results are presented for acquisition of the propeller tonals using real-world underwater acoustic data. Tonals estimated by third-order cyclic modulation bicoherence are more notable than the ones obtained by second-order cyclic modulation coherence due to latter's higher noise floor. Sparse recovery results show that frequencies of the prominent tonals can be obtained with sampling significantly below the Nyquist rate. The accurate estimation of tonal magnitudes, on the other hand, is challenging even with large number of compressive samples. Compressive sensing can be extended to solve underdetermined system of equations which appears in direction-of-arrival estimation with uniform linear arrays. An estimator is proposed based on the compressive beamformer for cyclostationary propeller noise. Its asymptotic bias is derived, which is inherited from the conventional beamformer when there are multiple sources. Squared asymptotic bias and the finite-sample variance, also derived explicitly, constitute the mean-squared error. Spectral averaging is suggested to mitigate this error by decreasing the adverse effect of the spatial Dirichlet kernel. For low signal-to-noise ratios, averaging enables the proposed estimator to outperform the methods that assume stationarity. This is achieved even under weak cyclostationarity, numerous closely-spaced sources and few sensors. The proposed methods are not only suitable for compressive sensing of propeller cavitation noise but also for general class of cyclostationary signals. Relevant research areas include but are not limited to communication, radar, acoustics and mechanical systems with applications such as spectrum sensing, modulation recognition, time difference of arrival estimation, time-frequency distributions, compressive detection and rolling element bearing fault diagnosis.
-
ÖgeQualitative microwave imaging in non-destructive testing and evaluation applications(Graduate School, 2023-04-05)Microwave imaging is an inspiring research topic in which the goal is to obtain constitutive properties of inaccessible targets using measurements of the scattered electric field or scattering parameters. The phrase microwave refers to the frequencies of the electromagnetic fields used in this technology, which can range from several hundred MHz to several hundred GHz. The wavelength of the fields allowed us to analyze the materials without causing any damage within this frequency bandwidth. Because of this characteristic, this research area has discovered various non-destructive testing and evaluation applications in applied sciences, such as biomedical imaging, moving target detection, food imaging, subsurface imaging, concealed weapon detection, and through-the-wall imaging. Instead of reconstructing the electrical parameters, one technique for dealing with microwave inverse scattering problems is to compute an indicator function that holds the information of the targets. These techniques are known as qualitative microwave imaging methods (Q-MWM), and they are typically thought to be linear and non-iterative techniques that are computationally less expensive than their quantitative counterparts. The most extensively researched Q-MWM representatives are truncated singular value decomposition (TSVD), linear sampling method (LSM), and factorization method (FM). The singular sources method (SSM) and the nearfield orthogonality sampling method (NOSM) are comparatively new, yet they have a promising future in microwave imaging. In the first part of the thesis, the problem of microwave imaging of an impedance cylinder is investigated using Newton's approach. To achieve this goal, the scattered field from a circular cylinder with homogenous impedance is determined for plane wave illumination. At this stage, the scattering configuration is considered to be in the form of a TMz scenario. In this scenario, the impedance cylinder is supposed to be infinite along the z-axis, and the electric fields are assumed to be parallel to the same axis. The incident plane wave is assumed to be decomposed into a summation of Bessel functions, whereas the scattered field is assumed to be expressed as a sum of Hankel functions, according to these assumptions. After that, the boundary conditions on the surface of the impedance cylinder are utilized in order to acquire the unknown coefficients in the scattered field. After then, during the inverse scattering phase, it is necessary to make estimates concerning the target's impedance as well as its radius. To achieve this aim, the scattered field of the impedance cylinder is collected at a number of frequencies on a single point surrounding the target. Following this, an initial value is given to both of the variables, and the evaluation of the scattered field that corresponds to these initial values occurs. To arrive at an estimate of the updated amount for each parameter, the difference in the scattered fields is first divided into a matrix. This matrix then contains the derivative of the scattered field with respect to the unknown variables. Then, both of the parameters are updated, and this procedure is repeated as many times as necessary until the difference between the measured and estimated fields falls below a certain threshold that has been established. As a result, we are able to derive an estimate of the impedance as well as the radius of the cylinder. According to the findings, the method that is now suggested is capable of reconstructing the unknown parameters using only a limited aperture and several frequency observations. In the second part of the thesis, we examined differential through-the-wall microwave imaging with several formulations of the TSVD method in a non-anechoic experiment. Past studies have used TSVD with a single transmitting/measuring antenna, whereas we show how to use it with a moving linear transmitting/measuring antenna array. Particularly, for repeated measurements, an averaging procedure is adopted. Three TSVD approaches are tested: TSVD on Contrast Source, TSVD on Contrast and multi frequency TSVD on Contrast. The dimension of the inverted matrix in TSVD on Contrast Source method is comparatively small. Following the solution of equations, a normalization scheme is suggested to eliminate the noise. TSVD on Contrast technique produces better reconstructions than TSVD on Contrast Source method because measured data for all excitations are inverted simultaneously. TSVD on Contrast, however, takes a long time than TSVD on Contrast Source because the inverted matrix becomes larger. Finally, in order to avoid further calibration simulations/measurements in multi frequency TSVD on Contrast, we use TSVD on Contrast solutions to obtain the calibration information. The contrasts are then computed for all frequencies and excitations at the same time. Thus, for multi frequency TSVD on Contrast, the inverted matrix is the largest, the accuracy is the best, and also the computational burden is the greatest. A metallic scatterer is placed behind a wall to evaluate the proposed techniques. The results demonstrate a trade-off between accuracy and computational time when selecting an appropriate inversion approach. Furthermore, each method's norm type selection is evaluated. In the third part of the thesis, the imaging of moving objects with Q-MWM is addressed. The necessity of background measurement is a troublesome aspect of Q-MWM. To avoid this, the total electric field collected at distinct time instants (say, Etotn, Etotl are the total electric fields measured at nth and lth time instants) are implemented to Q-MWM. Thus, the outcome of the Q-MWM can be considered to be the sum of the indicators at these time instants (i.e. Etotn-Etotl produces the differential indicator Inl=In+Il, where In, Il are the indicators at nth and lth time instants). An equation system is developed for indicator values at different time instants using this information for all possible time couples. Without performing any background measurements, the indicator of Q-MWM for each time frame is derived by solving this equation system. The proposed algorithm's performance is validated using 3D and 2D (both transverse magnetic (TM) and transverse electric (TE)) experimental measurements, which are done in a non-anechoic environment, for the LSM, which is an example of Q-MWM. In the fourth part of the thesis, SSM, a qualitative imaging method, is investigated for two-dimensional transverse magnetic electromagnetic (2D-TM EM) inverse scattering cases. Qualitative microwave imaging approaches allow for the rapid and accurate reconstruction of target shapes from scattered electric field measurements. This section's contribution can be stated as follows: (i) The SSM was originally introduced for the far-field scenario; here, we extend the SSM in the near field - inhomogeneous background configuration. Each stage of the extension (which involves an integral equation) is discussed using the linearity and reciprocity principles to provide physical insights. (ii) A relationship is established between the electrical properties of the scatterers and the SSM indicator. (iii) The suggested method is examined for monitoring hyperthermia treatment problems with a realistic breast model to evaluate the performance of SSM in real-world scenarios. The obtained results demonstrate that the SSM is capable of handling realistic breast phantoms for monitoring hyperthermia problems. In the fifth and last part of the thesis, a range-migration technique is presented for near-field microwave imaging using monostatic and bistatic measurement configurations. Calibration measurements are critical for enhancing the precision of both qualitative and quantitative microwave imaging. A calibration measurement should ideally be taken at each desired range (or depth) position in three-dimensional (3-D) near-field imaging, which can be time-consuming. The calibration effort can be reduced to a single measurement at a reference range position if the range behavior of the resolvent kernel of scattering can be predicted analytically. Analytical formulations for range-translation (or range-migration) are already commonly utilized in far-zone radar and acoustic imaging; nevertheless, their accuracy suffers dramatically in near-field situations. The magnitude and phase of the system point-spread function (PSF) are accurately estimated at any desired range position based on a single measurement of the PSF. The proposed migration is conducted in real space, but it can also be implemented with Fourier-domain (or k-space) inversion methods. It is used in simulation-based and experimental examples to confirm its performance and demonstrate its limitations with quantitative microwave holography.
-
ÖgeMoleküler haberleşme sistemlerinde alıcı kestirim yöntemleri(Lisansüstü Eğitim Enstitüsü, 2023-03-22)Nanoteknoloji atomik, moleküler veya makromoleküler seviyelerdeki teknolojilerin incelenmesi, geliştirilmesi ve uygulanması olarak tanımlanır. Nanoteknolojinin temel fonksiyonel birimi nanomakinelerdir. Nanomakineler birkaç nanometreden birkaç mikrometreye kadar değişen boyutlarda olabilen otonom makinelerdir. Nanomakineler algılama, hesaplama, haberleşme ve çalıştırma gibi basit görevleri yapabilen nano ölçekte bileşenlerden oluşmaktadır. Tek bir nanomakine, boyutu ve basit yapısından dolayı sadece sınırlı görevleri yerine getirebilmektedir. Öte yandan, bir kaç nanomakine daha karmaşık görevleri yerine getirebilmek için bir ağ üzerinden iş birliği yapabilir. Bu ağlar nano ağlar olarak adlandırılmaktadır. Nano ağ, nanomakinelerin birbirleri ile haberleşerek iş birliği yaptığı ve bu sayede sağlık ölçümü, akıllı ilaç gönderimi, nano ölçekteki çevrelerde biyolojik ve kimyasal saldırıların tespiti gibi daha karmaşık görevlerin yerine getirilebildiği bir sistemdir. Enerji, sağlık, çevre, gıda ve tekstil gibi birçok kullanım alanına sahip olan nanomakineler, karmaşık görevleri yerine getirebilmek için birbirleri ile haberleşmeye ihtiyaç duymaktadırlar. Nanomakineler için literatürde mekanik, akustik, elektromanyetik ve moleküler haberleşme sistemleri gibi birçok haberleşme sistemi önerilmiştir. Bu sistemlerden çoğu nano ölçekli rejim ile ilgili güç, boyut ve karmaşıklık kısıtlamaları sebebiyle uygulanamaz olarak tanımlanmıştır. Öte yandan moleküler haberleşme nanomakineler için en uygun çözüm olarak öngörülmüştür. Moleküler haberleşme, nanomakinelerin akışkan bir ortamda kimyasal işaretler veya moleküller kullanarak haberleşebildiği bir sistemdir. Moleküler haberleşmede gönderici nanomakine, bilgiyi moleküllere kodlayarak akışkan ortama gönderir. Ortamda yayılan bilgi molekülleri alıcı nanomakine ile etkileşime girer. Alıcı nanomakine bilgi moleküllerini sezerek bilgiyi çözer ve uygulamaya yönelik görevleri yerine getirir. Nanomakineler için iyon işaretleşmesi, aktif taşıma, bakteri tabanlı haberleşme ve difüzyon tabanlı haberleşme gibi çeşitli moleküler haberleşme sistemleri araştırılmıştır. Bu haberleşme sistemlerinden difüzyon tabanlı moleküler haberleşme enerji-verimli bir haberleşme sistemidir. Difüzyon tabanlı moleküler haberleşme, moleküllerin akışkan ortamda difüzyon ile pasif bir şekilde yayıldığı moleküler haberleşme sistemlerinden biridir. Difüzyon tabanlı moleküler haberleşme sistemlerinde, moleküllerin rastgele difüzyon yayılımından dolayı, klasik radyo frekansı (RF) tabanlı haberleşmeye göre semboller arası girişim daha fazladır. Bu da haberleşme sisteminin güvenirliğini azaltmaktadır. Bunun yanı sıra, nanomakineler nano ölçekli otonom makineler olduğundan, enerji, boyut, hesaplama kabiliyeti açısından belirli kısıtlamalara sahiptirler. Nanomakinelerin sahip olduğu bu kısıtlamalar ve difüzyon tabanlı moleküler haberleşme sistemlerindeki yüksek semboller arası girişimden dolayı, difüzyon tabanlı moleküler haberleşme sistemlerinde, güvenilir ve uygulanabilir bir haberleşme için düşük hesaplama karmaşıklığına ve yüksek bit hata oranı performansına sahip alıcı kestirim yöntemleri gerekmektedir. Bu tezde, difüzyon tabanlı moleküler haberleşme sistemlerinde model tabanlı ve yapay zeka tabanlı alıcı kestirim yöntemleri araştırılmaktadır. Bu kapsamda, difüzyon tabanlı moleküler haberleşme sistemi için literatürde önerilen alıcı kestirim yöntemleri verilmiş, model ve yapay zeka tabanlı alıcı kestirim yöntemleri önerilmiştir. Literatürde önerilen en büyük olabilirlik dizi kestirim yöntemi, göz ardı edilen ISI terimlerinin oluşturduğu DC yanlılık hesaba katılarak değiştirilmiş ve bit hata performansı açısından daha iyi olan Viterbi dal metrikleri elde edilmiştir. Değiştirilmiş Viterbi algoritmasının, literatürde önerilen klasik Viterbi algoritması ile aynı hesaplama karmaşıklığına sahip olduğu gösterilmiştir. Öte yandan, değiştirilmiş Viterbi algoritması, klasik Viterbi algoritmasına göre daha iyi bit hata oranı performansı göstermektedir. Değiştirilmiş Viterbi algoritması önerilen alıcı kestirim yöntemlerinin bit hata oranı performansları için temel ölçüt olarak kullanılmıştır. İşarete bağımlı gözlem gürültüsünün varyansı yerine bu varyansın istatistiksel ortalaması kullanılarak durağan kanallarda geçerli olan Wiener filtre difüzyon tabanlı moleküler haberleşme sistemi için bir alıcı kestirim yöntemi olarak önerilmiştir. Durağan ve doğrusal olmayan kanallarda kullanılabilen genişletilmiş Kalman filtre, difüzyon tabanlı moleküler haberleşme için ilk kez bir alıcı kestirim yöntemi olarak önerilmiştir. Literatürde önerilen en az ortalama kareler hatası (MMSE) alıcı kestirim yöntemi, Wiener ve Kalman filtrelerinin bit hata oranı (BER) performansları için temel ölçüt olarak kullanılmıştır. Önerilen alıcı kestirim yöntemlerinin BER performansları ayrıca literatürde önerilen uyarlamalı eşik ve sabit eşik kestirim yöntemleri ile karşılaştırılmıştır. Diğer yandan, hesaplama karmaşıklığı Viterbi algoritmasına göre daha düşük olan genişletilmiş Kalman filtrenin, alıcı ve verici arasındaki uzaklık azalırken yada vericiden gönderilen molekül sayısı yüksekken Viterbi algoritmasından daha iyi BER performansı gösterdiği gözlemlenmiştir. Dahası, genişletilmiş Kalman filtre, Viterbi algoritmasından daha az durum sayısına sahipken bile, Viterbi algoritmasından daha iyi BER performansı göstermiştir. Wiener filtre, uyarlamalı eşik değeri kestirim yönteminden daha iyi BER performansı gösterirken, belirli koşullar altında sabit eşik değeri kestirim yöntemi ile benzer BER performansına sahip olduğu görülmüştür. Wiener filtrenin daha düşük hesaplama karmaşıklığı ile MMSE kestirim yönteminden daha iyi bir BER performansına sahip olduğu gösterilmiştir. Dahası, belirli bir bit süresi için sadece iki uzunluklu sonlu dürtü yanıtlı (finit impulse response, FIR) Wiener filtre yeterli bir BER performansı göstermiştir. Öte yandan, genişletilmiş Kalman filtre, ortalama kare hatasını minimize eden diğer alıcı kestirim yöntemleri MMSE ve Wiener filtreden daha iyi BER performansı göstermiştir. Sonuç olarak, lineer hesaplama karmaşıklığına sahip Wiener filtre ve karesel hesaplama karmaşıklığa sahip Kalman filtre moleküler haberleşme için yüksek BER performansı gösteren alıcı kestirim yöntemleri olarak önerilmiştir. Model tabanlı alıcı kestirim yöntemlerinin bir çoğu alıcıda kanal durum bilgisine (channel state information, CSI) ihtiyaç duymaktadır. Moleküler haberleşme kanalında gözlem gürültüsü vericiden gönderilen işarete istatistiksel olarak bağımlıdır. Bundan dolayı moleküler haberleşme kanalının durumu zamanla değişmektedir ve kanal durum bilgisinin alıcıda kestirimi zordur. Ayrıca, kanal durum bilgisinin alıcıdaki kestiriminin hassasiyeti, model tabanlı alıcı kestirim yöntemlerinin performansını etkileyebilmektedir. Öte yandan, yapay zeka tabanlı alıcı kestirim yöntemleri, alıcıda kanal durum bilgisine (CSI) ihtiyaç duymamakta, sadece alıcıya gelen işareti kullanarak kestirim yapmaktadır. Makine öğrenmesi ve derin öğrenme modelleri alıcıda kanal durum bilgisine ihtiyaç duymadan difüzyon tabanlı moleküler haberleşme sistemi için yapay zeka tabanlı alıcı kestirim yöntemi olarak kullanılabilir. Bu tezde, karar ağacı, çok modelli rastgele örnekleme, karar ağacı ormanı ve uyarlamalı takviyeli sınıflandırma makine öğrenmesi algoritmaları difüzyon tabanlı moleküler haberleşme sistemi için önerilmiş ve BER performansları literatürde önerilen sabit eşik kestirim yöntemi ve ileri beslemeli tam bağlantılı yapay sinir ağı alıcı kestirim yöntemi ile karşılaştırılmıştır. Tek bir karar ağacı kullanıldığı zaman, yapay sinir ağı ve sabit eşik değeri kestirim yöntemlerinin karar ağacı algoritmasından daha iyi BER performansına sahip olduğu görülmüştür. Öte yandan, çok modelli rastgele örnekleme, karar ağacı ormanı ve uyarlamalı takviyeli sınıflandırma gruplama yöntemlerinin yapay sinir ağı ve sabit eşik değeri kestirim yöntemlerinden daha iyi BER performansına sahip olduğu görülmüştür.
-
ÖgeImaging in layered media(Graduate School, 2023-05-18)Detection and imaging of obstacles embedded in a stratified medium find a wide application area including underground imaging and through-wall imaging (TWI) as two-layered and three-layered medium applications, respectively. Despite the fact that both subsurface imaging and TWI have been a widespread area of research, in-wall imaging has remained a relatively untouched area. The literature on this topic is very limited indicating the difficulty level of the problem, that is, the main concern of this thesis. In addition to radar techniques and inversion algorithms, direct methods to probe a structure stand out as a field of study that many applications are clustered around. Direct methods include the multiple signal classification (MUSIC) method, the linear sampling method (LSM), the factorization method, the point source method, and several others. Also, it is proven that the reverse time migration (RTM) method is of use in imaging applications. In this thesis, the RTM method is applied to the single-frequency reconstruction of embedded obstacles in a wall to perform an introductory study for in-wall imaging. The aim is to determine the geometrical properties of an object embedded in a wall by the use of an information function provided via the RTM method. The method is based on the computation of that information function separately at each point on a reconstruction domain. It is defined as the correlation levels between the incident fields emitted from the sources and the back-propagation of the scattered field. Reporting the level of achievement of the RTM method in in-wall imaging based on a single test would be misleading. It should be tested under different circumstances. For this purpose, the problem is taken from a broader perspective in order to show and confirm the effectiveness of the method. Numerical experiments within a fundamental scenario are determined in a particular order to perform an essential Monte Carlo simulation. The method is tested for different material types, locations, and structural properties of the embedded object. Both dielectric and perfect electric conductor (PEC) scatterers are taken into consideration. In other words, both weak and strong scatterers are studied. Also, different acquisition lines, operating frequencies, and noise levels are used. A comparative study should be conducted to make an objective evaluation of the achievement level of the method. In line with this, the performance level in in-wall imaging is shown in comparison with free space imaging to make a proper evaluation of the feasibility level of the method. A mathematical model is proposed to make an overall quantitative evaluation. Thus, it will be possible to make a general conclusion. Inverse crime is a challenge to be avoided which may occur in inverse scattering applications. It can be briefly described by the philosophical norm as "two wrongs do not make a right" that implies trivial solutions and meaningless achievements. It could be a sneaky trap while acquiring the data synthetically. In order to avoid inverse crime, the problem must be modeled properly. It is verily known that one usual suspect in inverse crime is the computation of mathematical expressions. Those are mostly the integral representations of field quantities and the Green's functions. Precision and accuracy must always be guaranteed at an acceptable level since the computation step is the weakest link that may lead to occur an inverse crime.
-
ÖgeNew combined non-orthogonal multiple access techniques for wireless networks(Graduate School, 2023-03-23)Thanks to orthogonal multiple access (OMA) schemes, connectivity is provided to multiple users, different services, or diverse applications by allocating non-overlapping hence orthogonal resource blocks (time, frequency, and code) in today's networks. The orthogonal resource allocation urges the exploitation of higher frequency bands toward visible light to provide wider bandwidths that accommodate the massive connectivity and the tremendous data traffic volume offered by the next-generation mobile networks. As 5G technology becomes available worldwide and 6G technology appears on the horizon, the only characteristic remaining unchanged is the need for the wider bandwidths. However, spectrum as an inherently scarce resource is already inefficiently used and the use of higher frequency bands causes coverage problems even if sensitive coverage planning is accomplished. Hence new techniques enhancing spectral efficiency and enabling hyper-connectivity are required for future networks. Non-orthogonal multiple access (NOMA) constitutes a promising solution for spectral efficiency and hyper-connectivity by allowing multiple users simultaneously to share the same resource block. NOMA benefits from multiplexing the signals from different sources by the nature of the wireless medium, where superposition coding is performed at the transmitter side leading to inter-user interference on purpose. To distinguish the different users at the receiver, successive interference cancellation (SIC) is applied. NOMA can be implemented in both downlink (DL) and uplink (UL). In downlink NOMA (DL-NOMA), a base station (BS) transmits the superposition coded signal to the users, contrary to the uplink NOMA (UL-NOMA), where the information flow is from the users to the BS. Hence, DL-NOMA differs from UL-NOMA in terms of the received superposition coded signal. The received signal is only affected by the respective channel fading coefficient in DL-NOMA. On the other hand, in the UL-NOMA where the signals of users are multiplexed over the air, the received superposition coded signal is not only the weighted sum of users' signals but also includes the respective channel fading coefficients of all users. Therefore adaptive SIC techniques are required to avoid error floor characteristics in the performance curves of UL-NOMA. In this thesis, NOMA has been studied in combination with other emerging technologies such as cooperative communication, full-duplex (FD) technology, cognitive radio (CR), energy harvesting (EH), and transmit diversity to further enhance spectral and energy efficiencies as well as to provide massive connectivity in a fair manner. Cooperative communication provides diversity by creating a virtual antenna array with the help of an intermediate relaying node in the lack of multiple antennas. Higher data rates and diversity gains are obtained, furthermore, coverage can be extended to unserved or underserved areas. CR is another promising solution to the inefficient use of spectrum, where unlicensed (secondary) users (SUs) can access the licensed (primary) users (PUs) band via dynamic spectrum access techniques. In the thesis, a cooperative NOMA (C-NOMA) scheme adopting a cognitive butterfly network design is considered, where far user's transmitter-receiver pair is assumed as PU and the near user's transmitter-receiver pair along with a relay is considered as SU. Butterfly network design involves both UL-NOMA and DL-NOMA transmission steps. Transmission from transmitter to the relay is according to the UL-NOMA principles while relay conveys the superposition coded signal to the receivers according to the DL-NOMA principles. Since the transmitter of each user is close to the receiver of other user while its receiver is near the other user's transmitter, the information from the side short links is also exploited to improve SIC performance. Although UL-NOMA has rarely found a place in the NOMA literature, it is a part of the end-to-end connection between transmitter-receiver pairs in this study and to the best of our knowledge, this is the first study combining UL-NOMA and DL-NOMA in one system configuration to investigate bit error probability (BEP) performance. Since the power allocation is of most importance and has a direct impact on performance, power allocation for both UL-NOMA and DL-NOMA steps of the end-to-end connection is investigated and the end-to-end BEP analyses are provided. A similar C-NOMA scheme adopting cognitive butterfly network design is examined by assuming that the relay has the ability to harvest energy from the radio frequency (RF) signals in the environment as a difference. It is assumed that energy harvesting can be applied by relay according to two energy harvesting architectures namely power splitting and ideal receiver. BEP analyses for both users in each energy harvesting architecture are performed. This study fills another gap in the NOMA literature, since a combination of C-NOMA, EH, and CR has not been investigated in terms of BEP performance yet. For both architectures, the comparisons between reference OMA schemes are also provided for the same energy and spectral efficiency to show their superiority. Although C-NOMA has advantages, cooperation degrades the spectral efficiency, which is the most significant advantage of NOMA, due to the need for an extra time slot. The potential of NOMA can be fulfilled thanks to FD technology providing the ability of simultaneous transmission and reception to relaying node. In the thesis, an UL C-NOMA network adopting FD technology is considered in the presence of a direct link between BS and the far user. It is assumed that there are multiple uniformly distributed near users within a disc seeking the opportunity to access the frequency band and the distance between BS and the far user determines the diameter of the disc. Since NOMA is not effective for more than 2 users, a user pairing strategy is performed to select a near user, which has the ability to work in FD mode. To eliminate the error floor in the performance curves, adaptive decoding order, namely hybrid SIC, is applied and compared to two SIC schemes, namely, channel state information (CSI)-based and quality-of-service (QoS)-based SIC, to show its superiority. CSI-based SIC has been widely studied in the NOMA literature, where decoding order is defined according to the CSIs of the users. In QoS-based SIC, the near user can access the band if the QoS requirement of the far user is ensured, otherwise, OMA is adopted. It is shown by the outage performance curves that hybrid SIC as a combination of CSI-based and QoS-based SIC eliminates the error floor characteristic in UL-NOMA networks and is robust to the self-interference caused by FD mode contrary to the other two SIC schemes. The computational complexity is also investigated and compared to that of the half-duplex NOMA scheme. Instead of the virtual antenna array, multiple antennas can be located at the transmitter to improve capacity and reliability. In this thesis, a transmit diversity scheme, coordinate interleaved orthogonal design (CIOD), is applied to DL-NOMA. CIOD performs an interleaving procedure where coordinates of the modulated signals experience independent channel fading coefficients resulting in modulation diversity. Two cell-edge users are served by two BSs from adjacent cells, hence multiple antenna technique is implemented in a distributed manner. BSs simultaneously transmit desired signals to both users according to the CIOD transmission principle, a $2\times2$ CIOD-NOMA matrix is emerged where the desired signals of users are located at reverse diagonals. BEP analyses are provided for both users. It is shown that full diversity is obtained at full rate for both users and significant coding gains are achieved. Power allocation is also investigated in terms of the tradeoff between fairness and BEP performance. Satisfactory fairness performance is achieved in return for an acceptable BEP degradation. Furthermore, another transmit diversity scheme is studied in the scope of this thesis. A two-stage transmit antenna selection (TAS) scheme where one among multiple antennas is chosen at BS, is investigated for DL-NOMA. The selection is based on the adopted performance criteria, which is chosen as the sum rate maximization in this study. In the first stage, an antenna subset maximizing sum rate is selected. In the second stage, Jain's fairness indices are obtained according to the individual achievable rates of the users and the antenna maximizing the fairness index is selected. With this scheme, the sum rate is maximized while fairness is ensured. Although the sum rate maximization by considering user fairness is significant, a working system requires an acceptable error performance. In NOMA, the users' signals can be distinguished at the receiver side thanks to the differences in their received signal powers. As the power difference between the signals increases, it becomes easier to distinguish them, which improves the SIC performance. Therefore, in this study a power allocation scheme is also considered to fix the difference between power allocation parameters which otherwise tend to be equal with the increasing signal-to-noise ratio (SNR). The power allocation scheme involves the maximization of the sum rate under the constraint of a target data rate for the far user. NOMA is recognized as an enabler for the next generation of mobile networks thanks to allowing massive connectivity and enhancing spectral efficiency and fairness between users. In the scope of this thesis, NOMA has been combined with above-mentioned emerging technologies, and the proposed protocols have been investigated in terms of BEPs and/or outage probabilities to fill the gaps in the NOMA literature.