Matching Items (196)
Description
Recent advances in camera architectures and associated mathematical representations now enable compressive acquisition of images and videos at low data-rates. While most computer vision applications of today are composed of conventional cameras, which collect a large amount redundant data and power hungry embedded systems, which compress the collected data for further processing, compressive cameras offer the advantage of direct acquisition of data in compressed domain and hence readily promise to find applicability in computer vision, particularly in environments hampered by limited communication bandwidths. However, despite the significant progress in theory and methods of compressive sensing, little headway has been made in developing systems for such applications by exploiting the merits of compressive sensing. In such a setting, we consider the problem of activity recognition, which is an important inference problem in many security and surveillance applications. Since all successful activity recognition systems involve detection of human, followed by recognition, a potential fully functioning system motivated by compressive camera would involve the tracking of human, which requires the reconstruction of atleast the initial few frames to detect the human. Once the human is tracked, the recognition part of the system requires only the features to be extracted from the tracked sequences, which can be the reconstructed images or the compressed measurements of such sequences. However, it is desirable in resource constrained environments that these features be extracted from the compressive measurements without reconstruction. Motivated by this, in this thesis, we propose a framework for understanding activities as a non-linear dynamical system, and propose a robust, generalizable feature that can be extracted directly from the compressed measurements without reconstructing the original video frames. The proposed feature is termed recurrence texture and is motivated from recurrence analysis of non-linear dynamical systems. We show that it is possible to obtain discriminative features directly from the compressed stream and show its utility in recognition of activities at very low data rates.
ContributorsKulkarni, Kuldeep Sharad (Author) / Turaga, Pavan (Thesis advisor) / Spanias, Andreas (Committee member) / Frakes, David (Committee member) / Arizona State University (Publisher)
Created2012
Description
A novel strain sensing procedure using an optical scanning methodology and diffraction grating is explored. The motivation behind this study is due to uneven thermal strain distribution across semiconductor chips that are composed of varying materials. Due to the unique properties of the materials and the different coefficients of thermal expansion (CTE), one can expect the material that experiences the highest strain to be the most likely failure point of the chip. As such, there is a need for a strain sensing technique that offers a very high strain sensitivity, a high spatial resolution while simultaneously achieving a large field of view. This study goes through the optical setup as well as the evolution of the optical grating in an effort to improve the strain sensitivity of this setup.
ContributorsChen, George (Co-author) / Ma, Teng (Co-author) / Liang, Hanshuang (Co-author) / Song, Zeming (Co-author) / Nguyen, Hoa (Co-author) / Yu, Hongbin (Thesis director) / Jiang, Hanqing (Committee member) / Barrett, The Honors College (Contributor) / Electrical Engineering Program (Contributor)
Created2014-05
Description
The zinc oxide nanowires being grown are not developing properly and need to be fixed. In order to do this, the furnace equipment and experimental procedure must be tested until the results produced yield acceptable quality zinc oxide nanowires. After experimentation the nanowires were produced to an acceptable quality. With quality nanowires to experiment with, testing began to examine the effects of different thicknesses of aluminum dopants. Once doped and annealed, the wires were transferred to a substrate with a grid so contact points could be applied. However; the experiment was phased out once this step was half way complete due to the lab shifting to examine co-doping zinc oxide nanowires as explored in part two of this paper. The goal of co-doping zinc oxide film is to create an ideal p
type relationship for power generation, so this project focuses on altering the electrical properties of zinc oxide through doping that will allow more energy to be generated from the solar panels than current zinc oxide solar panels. The zinc oxide film doped with manganese was sputtered onto a silicon substrate. The experiment failed to create a co-doped sample because an x-ray photoelectron spectroscopy reading of the sample proved no nitrogen existed in the zinc oxide doped with manganese film. This experiment leads into this research teams work with co-doping, so instead of viewing this project as a failure it is seen as a learning experience. The research team is examining the results and creating new experiments to run to fix the problem. I currently work with my mentor Dr. Hongbin Yu and Seung Ho Ahn while doing research.
type relationship for power generation, so this project focuses on altering the electrical properties of zinc oxide through doping that will allow more energy to be generated from the solar panels than current zinc oxide solar panels. The zinc oxide film doped with manganese was sputtered onto a silicon substrate. The experiment failed to create a co-doped sample because an x-ray photoelectron spectroscopy reading of the sample proved no nitrogen existed in the zinc oxide doped with manganese film. This experiment leads into this research teams work with co-doping, so instead of viewing this project as a failure it is seen as a learning experience. The research team is examining the results and creating new experiments to run to fix the problem. I currently work with my mentor Dr. Hongbin Yu and Seung Ho Ahn while doing research.
ContributorsBull, David Sean (Author) / Yu, Hongbin (Thesis director) / Ahn, Seung Ho (Committee member) / Barrett, The Honors College (Contributor) / Chemical Engineering Program (Contributor)
Created2014-05
Description
Graphene, a one atomic thick planar sheet of carbon atoms, has a zero gap band structure with a linear dispersion relation. This unique property makes graphene a favorite for physicists and engineers, who are trying to understand the mechanism of charge transport in graphene and using it as channel material for field effect transistor (FET) beyond silicon. Therefore, an in-depth exploring of these electrical properties of graphene is urgent, which is the purpose of this dissertation. In this dissertation, the charge transport and quantum capacitance of graphene were studied. Firstly, the transport properties of back-gated graphene transistor covering by high dielectric medium were systematically studied. The gate efficiency increased by up to two orders of magnitude in the presence of a high top dielectric medium, but the mobility did not change significantly. The results strongly suggested that the previously reported top dielectric medium-induced charge transport properties of graphene FETs were possibly due to the increase of gate capacitance, rather than enhancement of carrier mobility. Secondly, a direct measurement of quantum capacitance of graphene was performed. The quantum capacitance displayed a non-zero minimum at the Dirac point and a linear increase on both sides of the minimum with relatively small slopes. The findings - which were not predicted by theory for ideal graphene - suggested that scattering from charged impurities also influences the quantum capacitance. The capacitances in aqueous solutions at different ionic concentrations were also measured, which strongly suggested that the longstanding puzzle about the interfacial capacitance in carbon-based electrodes had a quantum origin. Finally, the transport and quantum capacitance of epitaxial graphene were studied simultaneously, the quantum capacitance of epitaxial graphene was extracted, which was similar to that of exfoliated graphene near the Dirac Point, but exhibited a large sub-linear behavior at high carrier density. The self-consistent theory was found to provide a reasonable description of the transport data of the epitaxial graphene device, but a more complete theory was needed to explain both the transport and quantum capacitance data.
ContributorsXia, Jilin (Author) / Tao, N.J. (Thesis advisor) / Ferry, David (Committee member) / Thornton, Trevor (Committee member) / Tsui, Raymond (Committee member) / Yu, Hongbin (Committee member) / Arizona State University (Publisher)
Created2010
Description
Human activity recognition is the task of identifying a person’s movement from sensors in a wearable device, such as a smartphone, smartwatch, or a medical-grade device. A great method for this task is machine learning, which is the study of algorithms that learn and improve on their own with the help of massive amounts of useful data. These classification models can accurately classify activities with the time-series data from accelerometers and gyroscopes. A significant way to improve the accuracy of these machine learning models is preprocessing the data, essentially augmenting data to make the identification of each activity, or class, easier for the model. <br/>On this topic, this paper explains the design of SigNorm, a new web application which lets users conveniently transform time-series data and view the effects of those transformations in a code-free, browser-based user interface. The second and final section explains my take on a human activity recognition problem, which involves comparing a preprocessed dataset to an un-augmented one, and comparing the differences in accuracy using a one-dimensional convolutional neural network to make classifications.
ContributorsLi, Vincent (Author) / Turaga, Pavan (Thesis director) / Buman, Matthew (Committee member) / Computer Science and Engineering Program (Contributor) / Barrett, The Honors College (Contributor)
Created2021-05
Description
At present, the vast majority of human subjects with neurological disease are still diagnosed through in-person assessments and qualitative analysis of patient data. In this paper, we propose to use Topological Data Analysis (TDA) together with machine learning tools to automate the process of Parkinson’s disease classification and severity assessment. An automated, stable, and accurate method to evaluate Parkinson’s would be significant in streamlining diagnoses of patients and providing families more time for corrective measures. We propose a methodology which incorporates TDA into analyzing Parkinson’s disease postural shifts data through the representation of persistence images. Studying the topology of a system has proven to be invariant to small changes in data and has been shown to perform well in discrimination tasks. The contributions of the paper are twofold. We propose a method to 1) classify healthy patients from those afflicted by disease and 2) diagnose the severity of disease. We explore the use of the proposed method in an application involving a Parkinson’s disease dataset comprised of healthy-elderly, healthy-young and Parkinson’s disease patients.
ContributorsRahman, Farhan Nadir (Co-author) / Nawar, Afra (Co-author) / Turaga, Pavan (Thesis director) / Krishnamurthi, Narayanan (Committee member) / Electrical Engineering Program (Contributor) / Computer Science and Engineering Program (Contributor) / Barrett, The Honors College (Contributor)
Created2020-05
Description
This paper will primarily deal with obstacle detection and the benefits that radar technology provides as the primary interface. The concept that is being proposed involves using a non-industrialized radar to achieve similar results when trying to detect a present object. By being able to achieve a working radar detection system at a more general domain, the path to it becoming more universal accessible increases. This, in turn, will hopefully amplify the areas in which radar technology can be applied to and lead to great benefits universally. From the compiled data and the work that has been done to achieve a responsive radar, it is noted that the radar will provide an accurate reading in most conditions that it is introduced to. These conditions vary from range resolution aspects to various weather environments, as well as the visibility aspect. However, based on the results that were achieved, through various testing, there are still some areas in which radar technology needs to improve in, for it to be fully considered as the sole interface when it comes to obstacle detection and its integration into future technology like self-driving cars. Nevertheless, the capabilities of radar technology at this caliber is noted to be quite impressive and similar to other more expansive options that are available.
ContributorsMartinez, Johan (Author) / Yu, Hongbin (Thesis director) / Houghton, Todd (Committee member) / Electrical Engineering Program (Contributor, Contributor) / Barrett, The Honors College (Contributor)
Created2019-12
Description
Two nested capacitors can produce work if the electric fields are not aligned, and the purpose of this research was to explore the possibility of using that generation instead of DC motors. The work the capacitors produce is determined by the strength of the fields and materials that is composed of. The power density of the object is then determined by the volume. As the electric field increases in strength, the power increases, so to create a very strong internal field. The nested capacitors use a dielectric to prevent breakdown from the strength of the field. Additionally, as the nested capacitors decrease in size, their power density increases rapidly \u2014 becoming close to a dc motor's power density around the 500mm^2 size. When the result was simulated, it was discovered that the electric field was not contained to the dielectric and would result in sparking. Several other concerns would need to be addressed for this to become a viable solution.
ContributorsFryda, George Andrew (Author) / Singh, Anoop (Thesis director) / Yu, Hongbin (Committee member) / Mechanical and Aerospace Engineering Program (Contributor) / Barrett, The Honors College (Contributor)
Created2018-05
Description
The focus of this study was to address the problem of prohibitively expensive LiDARs currently being used in autonomous vehicles by analyzing the capabilities and shortcomings of affordable LiDARs as replacements. This involved the characterization of affordable LiDARs that are currently available on the market. The characterization of the LiDARs involved testing refresh rates, field of view, distance the sensors could detect, reflectivity, and power of the emitters. The four LiDARs examined in this study were the Scanse, RPLIDAR A2, LeddarTech Vu8, and LeddarTech M16. Of these low cost LiDAR options we find the two best options for use in affordable autonomous vehicle sensors to be the RPLIDAR A2 and the LeddarTech M16.
ContributorsMurphy, Thomas Joseph (Co-author) / Gamal, Eltohamy (Co-author) / Yu, Hongbin (Thesis director) / Houghton, Todd (Committee member) / Electrical Engineering Program (Contributor) / W.P. Carey School of Business (Contributor) / Barrett, The Honors College (Contributor)
Created2018-05
Description
This research report investigates the feasibility of using RFID in Traffic Sign Recognition (TSR) Systems for autonomous vehicles, specifically driver-less cars. Driver-less cars are becoming more prominent in society but must be designed to integrate with the current transportation infrastructure. Current research in TSR systems use image processing as well as LIDAR to identify traffic signs, yet these are highly dependent on lighting conditions, camera quality and sign visibility. The read rates of current TSR systems in literature are approximately 96 percent. The usage of RFID in TSR systems can improve the performance of traditional TSR systems. An RFID TSR was designed for the Autonomous Pheeno Test-bed at the Arizona State University (ASU) Autonomous Collective Systems (ACS) Laboratory. The system was tested with varying parameters to see the effect of the parameters on the read rate. It was found that high reader strength and low tag distance had a maximum read rate of 96.3 percent, which is comparable to existing literature. It was proven that an RFID TSR can perform as well as traditional TSR systems, and has the capacity to improve accuracy when used alongside RGB cameras and LIDAR.
ContributorsMendoza, Madilyn Kido (Author) / Berman, Spring (Thesis director) / Yu, Hongbin (Committee member) / Mechanical and Aerospace Engineering Program (Contributor) / Barrett, The Honors College (Contributor)
Created2018-05