The past decade witnessed the rapid development of III-nitride light-emitting diodes (LEDs) and laser diodes (LDs), for smart lighting, visible-light communication (VLC), optical storage, and internet-of-things. Recent studies suggested that the GaN-based LDs, which is free from efficiency droop, outperform LEDs as a viable high-power light source. Conventionally, the InGaN-based LDs are grown on polar, c-plane GaN substrates. However, a relatively low differential gain limited the device performance due to a significant polarization field in the active region. Therefore, the LDs grown on nonpolar m-plane and semipolar (2021)-plane GaN substrates are posed to deliver high-efficiency owing to the entirely or partially eliminated polarization field. To date, the smart lighting and VLC functionalities have been demonstrated based on discrete devices, such as LDs, transverse-transmission modulators, and waveguide photodetectors. The integration of III-nitride photonic components, including the light emitter, modulator, absorber, amplifier, and photodetector, towards the realization of III-nitride photonic integrated circuit (PIC) offers the advantages of small-footprint, high-speed, and low power consumption, which has yet to be investigated. This dissertation presents the design, fabrication, and characterization of the multi-section InGaN laser diodes with integrated functionalities on semipolar (2021)-plane GaN substrates for enabling such photonic integration. The blue-emitting integrated waveguide modulator-laser diode (IWM-LD) exhibits a high modulation efficiency of 2.68 dB/V. A large extinction ratio of 11.3 dB is measured in the violet-emitting IWM-LD. Utilizing an integrated absorber, a high optical power (250mW), droop-free, speckle-free, and large modulation bandwidth (560MHz) blue-emitting superluminescent diode is reported. An integrated short-wavelength semiconductor optical amplifier with the laser diode at ~404 nm is demonstrated with a large gain of 5.32 dB at 6 V. A high-performance waveguide photodetector integrated LD at 405 nm sharing the single active region is presented, showing a significant large modulation bandwidth of 230 MHz. Thus these seamlessly integrated elements enable photonic IC at the visible wavelength for many important applications, such as smart lighting and display, optical communication, switching, clocking, and interconnect. The findings are therefore significant in developing an energy-saving platform technology that powers up human activities in a safe, health- and environmental-friendly manner.

No description available.

Optoelectronics in the deep-ultraviolet (DUV) regime is still a growing research field that requires significant effort to understand the material properties and optimize the device structures to realize efficient DUV devices. Aluminum gallium nitride (AlGaN) is perhaps the most studied semiconductor to replace the environmentally hazardous mercury lamps; however, the external quantum efficiency of AlGaN based DUV devices is insufficient to replace the existing old-fashioned mercury UV lamps. Despite the tunability in the bandgap of AlGaN, the excessive strain accumulation associated with increased alloying of Al in AlGaN and the poor dopant activation due to the relatively large ionization energy of the donors and acceptors are not favorable for realizing efficient DUV emitters. In addition, the crossover among the light hole, heavy hole and split-off bands in the valance band for Al-rich AlGaN suppresses the transverse-electric polarization, which further worsens the external quantum efficiency. Furthermore, for DUV photodetection, commercially available Si-photodetectors suffer from poor responsivity for wavelengths shorter than 400 nm in contrast to the visible spectrum. Hence, the-state-of-art photodetectors in the DUV regime also need a significant upgrade, particularly for high-speed applications. Firstly, we utilized the high quantum confinement in plasma-assisted MBE grown ultrathin GaN QDisks to realize DUV (λ ≈ 260 nm) emission using a binary compound (GaN) in contrast to conventionally used ternary compound (AlGaN). More importantly, we experimentally demonstrated TE-dominant DUV emission, unlike Al-rich AlGaN, illustrating a unique pathway for realizing efficient DUV vertical emitters. Secondly, inspired by the light manipulation technique practiced in nature, we presented iridocytes on giant clams (Tridacna maxima), known for their symbiotic relationship with algae as a color downconverting material for DUV photodetection. Investigating the structural and optical properties of iridocytes found in Tridacna maxima, we established a robust UV communication allowing the data transfer rate of 100 Mbit/s within the forward error correction limit for modulated 375 nm-laser diode. Using a similar matrix implemented to 375 nm-laser, with high-power UV-C LED (λ ≈ 278 nm), we could establish an optical wireless communication that could allow a data-transmission rate of tens of Mbit/s within the forward error correction limit.

Tato diplomova prace se zabyva prpravou a charakterizac GaN nanokrystalu s kovovym jadrem. V teoreticke casti teto prace je predstaven material GaN se svymi vlastnos- tmi a aplikacemi. Dale jsou uvedeny substraty pro rust a jednotlive mechanismy rustu GaN nanokrystalu. V dalsm jsou popsany kovove nanocastice a jejich opticke vlastnosti umoznujc zesilovan fotoluminiscence na zaklade interakce plasmonu a GaN. Experi- mentaln cast se zabyva prpravou GaN nanokrystalu s Ag jadrem ve ctyrech krocch. Prvne jsou Ag nanocastice naneseny na substrat Si(111). Nasledne se nechaj zoxidovat. Tretm krokem je depozice Ga a poslednm je nitridace. Jednotlive kroky byly opti- malizovany a analyzovany ruznymi metodami, jako je XPS, SEM, fotoluminiscence a Ramanova spektroskopie.

The work deals with the generation of ECG signals using generative adversarial networks (GAN). It examines in detail the basics of artificial neural networks and the principles of their operation. It theoretically describes the use and operation and the most common types of failures of generative adversarial networks. In this work, a general procedure of signal preprocessing suitable for GAN training was derived, which was used to compile a database. In this work, a total of 3 different GAN models were designed and implemented. The results of the models were visually displayed and analyzed in detail. Finally, the work comments on the achieved results and suggests further research direction of methods dealing with the generation of ECG signals.

Generative Adversarial Networks (GANs) have attracted much attention because of their ability to learn high-dimensional, realistic data distributions. In the field of medical imaging, they can be used to augment the often small image sets available. In this way, for example, the training of image classification or segmentation models can be improved to support clinical decision making. GANs can be distinguished according to their input. While Noise-to-Image GANs synthesize new images from a random noise vector, Image-To-Image GANs translate a given image into another domain. In this study, it is investigated if the performance of a Noise-To-Image GAN, defined by its generated output quality and diversity, can be improved by using elements of a previously trained Image-To-Image GAN within its training. The data used consists of paired T1- and T2-weighted MR brain images. With the objective of generating additional T1-weighted images, a hybrid model (Hybrid GAN) is implemented that combines elements of a Deep Convolutional GAN (DCGAN) as a Noise-To-Image GAN and a Pix2Pix as an Image-To-Image GAN. Thereby, starting from the dependency of an input image, the model is gradually converted into a Noise-to-Image GAN. Performance is evaluated by the use of an independent classifier that estimates the divergence between the generative output distribution and the real data distribution. When comparing the Hybrid GAN performance with the DCGAN baseline, no improvement, neither in the quality nor in the diversity of the generated images, could be observed. Consequently, it could not be shown that the performance of a Noise-To-Image GAN is improved by using elements of a previously trained Image-To-Image GAN within its training.

In this thesis we investigate the use of GANs for texture enhancement. To achievethis, we have studied if synthetic satellite images generated by GANs will improvethe texture in satellite-based 3D maps. We investigate two GANs; SRGAN and pix2pix. SRGAN increases the pixelresolution of the satellite images by generating upsampled images from low resolutionimages. As for pip2pix, the GAN performs image-to-image translation bytranslating a source image to a target image, without changing the pixel resolution. We trained the GANs in two different approaches, named SAT-to-AER andSAT-to-AER-3D, where SAT, AER and AER-3D are different datasets provided bythe company Vricon. In the first approach, aerial images were used as groundtruth and in the second approach, rendered images from an aerial-based 3D mapwere used as ground truth. The procedure of enhancing the texture in a satellite-based 3D map was dividedin two steps; the generation of synthetic satellite images and the re-texturingof the 3D map. Synthetic satellite images generated by two SRGAN models andone pix2pix model were used for the re-texturing. The best results were presentedusing SRGAN in the SAT-to-AER approach, in where the re-textured 3Dmap had enhanced structures and an increased perceived quality. SRGAN alsopresented a good result in the SAT-to-AER-3D approach, where the re-textured3D map had changed color distribution and the road markers were easier to distinguishfrom the ground. The images generated by the pix2pix model presentedthe worst result. As for the SAT-to-AER approach, even though the syntheticsatellite images generated by pix2pix were somewhat enhanced and containedless noise, they had no significant impact in the re-texturing. In the SAT-to-AER-3D approach, none of the investigated models based on the pix2pix frameworkpresented any successful results. We concluded that GANs can be used as a texture enhancer using both aerialimages and images rendered from an aerial-based 3D map as ground truth. Theuse of GANs as a texture enhancer have great potential and have several interestingareas for future works.

The non-isolated POL converters are widely used in computers, telecommunication systems, portable electronics, and many other applications. These converters are usually constructed using discrete components, and operated at a lower frequency around 200 ~ 600 kHz to achieve a decent efficiency at the middle of 80's%. The passive components, such as inductors and capacitors, are bulky, and they occupy a considerable foot-print. As the power demands increase for POL converters and the limited real estate of the mother board, the POL converters must be made significantly smaller than what they have demonstrated to date. To achieve these goals, two things have to happen simultaneously. The first is a significant increase in the switching frequency to reduce the size and weight of the inductors and capacitors. The second is to integrate passive components, especially magnetics, with active components to realize the needed power density. Today, this concept has been demonstrated at a level less than 5A and a power density around 300-700W/in3 by using silicon-based power semiconductors. This might address the need of small hand-held equipment such as PDAs and smart phones. However, it is far from meeting the needs for applications, such as netbook, notebook, desk-top and server applications where tens and hundreds of amperes are needed. After 30 years of silicon MOSFET development, the silicon has approached its theoretical limits. The recently emerged GaN transistors as a possible candidate to replace silicon devices in various power conversion applications. GaN devices are high electron mobility transistors (HEMT) and have higher band-gap, higher electron mobility, and higher electron velocity than silicon devices, and offer the potential benefits for high frequency power conversions. By implementing the GaN device, it is possible to build the POL converter that can achieve high frequency, high power density, and high efficiency at the same time. GaN technology is in its early stage; however, its significant gains are projected in the future. The first generation GaN devices can outperform the state-of-the-art silicon devices with superior FOM and packaging. The objective of this work is to explore the design of high frequency, high power density 12 V input POL modules with GaN devices and the 3D integration technique. This work discusses the fundamental differences between the enhancement mode and depletion mode GaN transistors, the effect of parasitics on the performance of the high frequency GaN POL, the 3D technique to integrate the active layer with LTCC magnetic substrate, and the thermal design of a high density module using advanced substrates with improved thermal conductivity. The hardware demonstrators are two 12 V to 1.2 V highly integrated 3D POL modules, the single phase 10 A module and two phase 20 A module, all built with depletion mode GaN transistors and low profile LTCC inductors. / Master of Science

Telehealth is an online health care system that is extensively used in the current pandemic situation. Our proposed technique is considered a fog computing-based attack detection architecture to protect IoT Telehealth Networks. As for IoT Telehealth Networks, the sensor/actuator edge devices are considered the weakest link in the IoT system and are obvious targets of attacks such as botnet attacks. In this thesis, we introduce a novel framework that employs several machine learning and data analysis techniques to detect those attacks. We evaluate the effectiveness of the proposed framework using two publicly available datasets from real-world scenarios. These datasets contain a variety of attacks with different characteristics. The robustness of the proposed framework and its ability, to detect and distinguish between the existing IoT attacks that are tested by combining the two datasets for cross-evaluation. This combination is based on a novel technique for generating supplementary data instances, which employs GAN (generative adversarial networks) for data augmentation and to ensure that the number of samples and features are balanced. / Graduate

The old methods used for images inpainting of the Depth Image Based Rendering (DIBR) process are inefficient in producing high-quality virtual views from captured data. From the viewpoint of the original image, the generated data’s structure seems less distorted in the virtual view obtained by translation but when then the virtual view involves rotation, gaps and missing spaces become visible in the DIBR generated data. The typical approaches for filling the disocclusion tend to be slow, inefficient, and inaccurate. In this project, a modern technique Generative Adversarial Network (GAN) is used to fill the disocclusion. GAN consists of two or more neural networks that compete against each other and get trained. This study result shows that GAN can inpaint the disocclusion with a consistency of the structure. Additionally, another method (Filling) is used to enhance the quality of GAN and DIBR images. The statistical evaluation of results shows that GAN and filling method enhance the quality of DIBR images.