教学课程资源库 更多>>

关键词： Physiology   Architecture   Atmospheric sciences   Computer science   Energy  

摘要： Thermal comfort measurement in an indoor environment is essential as it impactspeople’s health and productivity. Several indices, such as effective temperature, heat stressindex, Oxford index, and wet bulb globe temperature, have been used to measure humanthermal comfort quantitatively over the past decades. Among other indices, severalinternational standards, such as ASHRAE 55-2017, ISO7730, EN 16798-1 and EN 16798-2, worldwide accept predicted mean vote (PMV) as an indicator of thermal comfortconditions, as it considers all key environmental and personal factors influencing humanthermal comfort. While several devices are available and reported both commercially andin the literature that measures PMV, most of these devices are very bulky and only takeenvironmental parameters into account. The focus of this thesis is to develop a CMOScompatible MEMS human thermal comfort sensing (HTCS) system based on PMV indexand apply it to indoor heating ventilation and air conditioning (HVAC) systems. First, we developed a PMV based micro HTCS system with multiple MEMS *** system comprised of a wireless multi-sensor (CMOS MEMS air velocity, MEMS RH,and MEMS air temperature) module to measure three environmental parameters and asmartphone App with a novel motion analytics algorithm (for estimation of metabolic rate) and personal factors (clothing insulation) input. While the majority of the reported HTCSsystems are bulky and take only environmental parameters into consideration, our developed system utilizes MEMS technology for sensors’ fabrication and measures theessential environmental and human factors involved to compute human thermal comfort. Then, to reduce both the size and cost of the HTCS system, we developed a single-chipfabrication process for CMOS compatible MEMS temperature, RH, and highly sensitiveair velocity sensors. Our multi-sensor chip (MSC) has two merits. First, it utilizes a low-cost3-mask fabrication process to fabricate temperature, RH, and TMCV se

关键词： Computer science   Oncology   Medical imaging   Artificial intelligence  

摘要： Breast cancer frequently occurs in women over the world. It was one of the most serious diseases and the second common cancer among women in 2019. The survival rate of stages 0 and 1 of breast cancer is closed to 100%. It is urgent to develop an approach that can detect breast cancer in the early stages. Breast ultrasound (BUS) imaging is low-cost, portable, and effective; therefore, it becomes the most crucial approach for breast cancer diagnosis. However, BUS images are of poor quality, low contrast, and uncertain. The computer-aided diagnosis (CAD) system is developed for breast cancer to prevent misdiagnosis. There have been many types of research for BUS image segmentation based on classic machine learning and computer vision methods, e.g., clustering methods, thresholding methods, level set, active contour, and graph cut. Since deep neural networks have been widely utilized in nature image semantic segmentation and achieved good results, deep learning approaches are also applied to BUS image segmentation. However, the previous methods still suffer some shortcomings. Firstly, the previous non-deep learning approaches highly depend on the manually selected features, such as texture, frequency, and intensity. Secondly, the previous deep learning approaches do not solve the uncertainty and noise in BUS images and deep learning architectures. Meanwhile, the previous methods also do not involve context information such as medical knowledge about breast cancer. In this work, three approaches are proposed to measure and reduce uncertainty and noise in deep neural networks. Also, three approaches are designed to involve medical knowledge and long-range distance context information in machine learning algorithms. The proposed methods are applied to breast ultrasound image segmentation. In the first part, three fuzzy uncertainty reduction architectures are designed to measure the uncertainty degree for pixels and channels in the convolutional feature maps. Then, medical

关键词： Sustainability   Energy   Petroleum engineering   Fluid mechanics   Computer science  

摘要： We improve an advanced fully-implicit parallel framework for reservoir simulation with the application of discretization and linearization methods. The multipoint flux approximation (MPFA) scheme is used with full tensor permeability and unstructured grids. Moreover, the main challenge in reservoir simulation is to have a complex phase behavior, as it leads to high nonlinearity. Operator-based linearization (OBL) has been introduced lately which simplifies residual & Jacobian assembly and also minimizes computational time furthermore. In OBL, the mass-based formulation is implemented in a fully implicit scheme to assure the constancy of solutions and flexibility of the framework for general-purpose reservoir simulation. In this work, we illustrate a complete investigation of the accuracy and efficiency of the OBL. There are three new points in this simulator first is the advanced parallel framework. We develop an advancing parallel framework for general-purpose reservoir simulation; it allocates a huge workload to multiple processors on a cluster. In this work, we use a multipoint linearization scheme that can deliver accurate, robust, and convergent solutions for reservoir simulation. We have a mimetic finite difference scheme (MFD) method and multipoint flux approximation method, which help in enhancing the simulation accuracy and efficiency. Besides, we compare the numerical solution with the analytical solution by running simulations with different resolutions. Moreover, we present several benchmark cases to thoroughly determine the accuracy, convergence, and robustness of the framework. We investigate the effect of grid resolution on simulation results by running several models with different grid sizes. We also investigate the robustness of the OBL by running simulations with different oil viscosities. Moreover, we run multi filed cases to illustrate the performance of the OBL method for reservoir simulations. Finally, MPFA-O helps to guarantee convergence sol

关键词： Aerospace engineering   Computer science  

摘要： Future LEO missions consisting of tens to thousands of satellites will be well-positioned to benefit from technologies to automate mission operations and utilize atmospheric interactions to perform station-keeping. This dissertation aims to demonstrate novel technical methods for both differential drag and spacecraft autonomy that minimizes overall system impacts while remaining straightforward to implement, use, and test. Two major areas of work are presented. In the first, a novel strategy for differential drag control using linearized attitude-orbit dynamics allows for the design of differential drag controllers that utilize small attitude variations. This control strategy is further examined and extended to deal with uncertainties in atmospheric density using a desensitized optimal control approach. The next area of work considers the applicability of deep reinforcement learning (DRL) to address challenges in spacecraft operations in a safe, performant, and scale-able manner. Frameworks and strategies for considering day-to-day spacecraft operations tasks as Markov Decision Processes are presented and discussed, culminating in the creation of several benchmark problems for spacecraft operations. The performance of DRL algorithms on each of these benchmark problems is presented and analyzed, demonstrating performance improvements against heuristic and black-box optimization approaches inspired by other strategies found in the literature. The synergy of both approaches is further demonstrated on a representative challenge consisting of a spacecraft conducting science operations in Low Earth Orbit (LEO) while phasing using differential drag. DRL agents are shown to successfully learn to sequence mission activities, phasing campaigns, and health management tasks using novel strategies learned by experience alone.

关键词： Computer science   Medical imaging   Acoustics   Remote sensing  

摘要： Spectrograms extract frequency components from a signal. Spectrograms have been in use for a long time mainly to analyze frequency components in audio signals. Typically, these audio signals have a very high sampling rate, various frequency components and high frequency variability with time. Vital signs on other hand have very low sampling rate with no frequency variability. This work explores if spectrograms can be used to analyze and recognize patterns from vital signs signals. As mentioned above, spectrograms deal with frequencies. More the variability of frequency, better the patterns emerge when spectrograms are applied on the signals. As vital signs lack the frequency variability, this work uses frequency modulation to introduce frequency variability. The amplitude of vital signs signal is encoded as frequency of another signal on which spectrogram is applied to understand the frequency variability which is directly proportional to the amplitude of the vital signs signal. Then, the generated spectrograms are given as an input to convolutional neural networks to extract features and classify the spectrograms. The efficacy of this method is tested on 4 different datasets for prediction and detection. The proposed method was able to achieve an accuracy of 91.55% and an AUC of 0.92 for prediction task and had better overall precision, recall, F1-score compared to the baseline method. It also achieved an accuracy of 91.67 and AUC of 0.92 on classification or detection task and also it performed better than baseline method in all the metrics such as precision, recall, F1-score for detection.

关键词： Computer science   Information technology   Artificial intelligence  

摘要： The introduction of Artificial Intelligence (AI) into the Cybersecurity domain impacts the cybersecurity professional, having massive ramifications on their future. Understanding the impact of AI requires the identification of manual tasks conducted today that can be automated using Artificial Intelligence. From AI identification, an understanding occurs, allowing for the future impact measurement on the cybersecurity professional’s duties. The reason for this study was to explore the identification of cybersecurity professional's daily manual tasks available for AI automation. The investigator used current scholarly and practitioner literature that was reviewed, analyzed, and summarized. This study used a qualitative exploratory design to collect data through Zoom non-visual interviews using semi-structured interview questions. In addition, the coding used was a non-hierarchical flat coding. Five major themes emerged from the data. The 5 major themes were (a) National Initiative for Cybersecurity Education (NICE) framework workforce categories; (b) rescoping of cybersecurity professional duties; (c) improved algorithms, data sets, and hardware; (d) individual trust in AI automation decisions; and (e) creative thinking is not automatable.

关键词： Materials science   Thermodynamics   Energy   Computer science   Artificial intelligence   Electrical engineering  

摘要： Active thermography is a thermal imaging technique that is mainly used in nondestructive testing. Invisible defects in materials, such as cracks or voids, can be detected by actively heating the surface of the material and using a thermal imaging camera. The diffusive nature of heat propagation causes blurring in the acquired thermal images and thus imposes a spatial resolution limit for closely spaced defects. Conventional thermographic techniques reach their spatial resolution limits at aspect ratios (distance between two defects: depth of defects) of 2:1 to 1:1. To increase the spatial resolution, so-called super resolution methods for laser thermography are developed and studied. The used high-power lasers allow spatially and temporally structured heating. Structured illumination requires multiple measurements to investigate the whole sample surface, but allows a significant increase of the spatial resolution, as already shown in optical microscopy. In addition, super resolution laser thermography, similar to optical super resolution, applies deconvolution algorithms that use the thermal point spread function. The underlying ill-posed inverse problem for defect reconstruction is solved using sparsity regularization in post-processing to promote sparse solutions, since defects usually appear sparsely. Moreover, the used l2,1-regularization favors pixels that are active over multiple measurements ("joint sparsity"). The focus in the dissertation is on blind illumination, where the inverse problem is solved without knowing the position of the illumination. To avoid a manual choice of regularization parameters and to achieve higher convergence rates for in-situ, high-resolution laser thermography, a suitable deep unfolding network is developed and studied within the doctoral thesis. A total of three different test specimens are examined, two of them made of structural steel (S235JR) with surface anomalies, the other one is additively manufactured from stainless stee

关键词： Medical imaging   Computer science   Artificial intelligence  

关键词： Remote sensing   Computer science   Artificial intelligence   Civil engineering  

摘要： In recent decades, due to global warming and climate change, we have witnessed ever-growing occurrences of natural disasters such as flooding, tornadoes, earthquakes, and wildfires. As such, it is more important than ever to provide emergency response personnel with accurate and timely information for their effective responses to crises. Among the variety of information which is needed for emergency responses and managements, it is vital that response personnel get informed in a timely manner as to where and how severely a building is damaged so that rescue efforts can be the most effective. However, challenges remain despite a great deal of effort in the field of image classification for disaster response. In this work, a promising approach based on deep learning is proposed to detect damaged buildings on high-resolution satellite imagery. By utilizing generic data augmentation, the proposed approach overcame the problem of limited training data popular in many remote sensing applications. A fine-tuning strategy is proposed to use transfer learning with a pretrained model for the task of interest. The experiments with imagery of Port-au-Prince, Haiti demonstrated the proposed method is effective when training data is limited. The Convolutional Neural Network (CNN) model with augmented training data can achieve 83% accuracy in detecting damaged buildings, greatly improved from 53% with the original training data. Future work would be focused on exploring automated methods to acquire larger training datasets and model generalization by investigating more robust data augmentation techniques.

关键词： Software   Embedded systems   Artificial intelligence   Decision making   Taxonomy   Algorithms   Validation studies   Computer science  

摘要： Context:Companies are increasingly collecting data from all possible sources to extract insights that help in data-driven decision-making. Increased data volume, variety, and velocity and the impact of poor quality data on the development of data products are leading companies to look for an improved data management approach that can accelerate the development of high-quality data products. Further, AI is being applied in a growing number of fields, and thus it is evolving as a horizontal technology. Consequently, AI components are increasingly been integrated into embedded systems along with electronics and software. We refer to these systems as AI-enhanced embedded systems. Given the strong dependence of AI on data, this expansion also creates a new space for applying data management techniques. Objective:The overall goal of this thesis is to empirically identify the data management challenges encountered during the development and maintenance of AI-enhanced embedded systems, propose an improved data management approach and empirically validate the proposed approach. Method:To achieve the goal, we conducted this research in close collaboration with Software Center companies using a combination of different empirical research methods: case studies, literature reviews, and action research. Results and conclusions:This research provides five main results. First, it identifies key data management challenges specific to Deep Learning models developed at embedded system companies. Second, it examines the practices such as DataOps and data pipelines that help to address data management challenges. We observed that DataOps is the best data management practice that improves the data quality and reduces the time to develop data products. The data pipeline is the critical component of DataOps that manages the data life cycle activities. The study also provides the potential faults at each step of the data pipeline and the corresponding mitigation strategies. Finally, the dat