12023 research outputs found
Sort by
Periods of Feynman Diagrams
Get PDFWe study differential equations for Feynman amplitudes and show that the corresponding D-module is isomorphic to a GKZ D-modules. We show that the sheaf of solutions to the D-module is isomorphic to a certain relative homology and that the amplitudes are periods of a relative motive. Using these ideas, we develop a method of regularization which specializes to dimensional regularization and analytic regularization.</p
Proximal-Field Radiation Sensors for Dynamically Controllable and Self-Correcting Integrated Radiators
Get PDFOne of the major challenges in the design of integrated radiators at mm-wave frequencies is the generation of surface waves in the dielectric substrate by the on-chip antennas. Since dielectric substrates are excellent surface waveguides with a fundamental mode with no cutoff frequency, there is always some energy trapped in them due to the surface waves and the excited substrate modes. This phenomenon is a significant cause of reduced radiation efficiency for mm-wave integrated radiators. However, in this thesis, we use this as an opportunity. We show that the excited substrate modes in the dielectric substrate of an integrated antenna contain valuable information regarding its far-field radiation properties. We introduce Proximal-Field Radiation Sensors (PFRS) as a number of small sensing antennas that are placed strategically on the same substrate as the integrated antenna and measure electromagnetic waves in its immediate proximity. These sensors extract the existing information in the substrate modes and use it to predict the far-field radiation properties of the integrated antenna in real-time based on in-situ measurements in the close proximity of the antennas, without any need to use additional test equipment and without removing the antenna from its operating environment or interfering with its operation in a wireless system. In other words, PFRS enables self-calibration, self-correction, and self-monitoring of the performance of the integrated antennas. Design intuition and a variety of data processing schemes for these sensors are discussed. Two proof-of-concept prototypes are fabricated on printed circuit board (PCB) and integrated circuit (IC) and both verify PFRS capabilities in prediction of radiation properties solely based on in-situ measurements.
Dynamically controllable integrated radiators would significantly benefit from PFRS, These radiators are capable of controlling their radiation parameters such as polarization and beam steering angle through their actuators and control units. In these cases, PFRS serves as a tool for real-time monitoring of their radiation parameters, so that without direct measurement of the far-field properties through bulky equipment the required information for the control units and the actuators are provided.
Dynamically controllable integrated radiators can be designed using the additional design space provided by Multi-Port Driven (MPD) radiator methodology. After a review of advantages of MPD design over the traditional single-port design, we show that a slot-based MPD radiator would have the additional advantage of reduced exclusive use area compared to the original wire-based MPD radiator, through demonstration of a 134.5-GHz integrated slot-based MPD radiator with a measured single-element EIRP of +6.0 dBm and a total radiated power of -1.3 dBm.
We discuss how MPD methodology enables the new concept of Dynamic Polarization Control, as a method to ensure polarization matching of the transmitter antenna to the receiver antenna, regardless of the polarization and orientation of the receiver antenna in space. A DPC antenna design using the MPD methodology is described and a 105.5-GHz 2x1 integrated DPC radiator array with a maximum EIRP of +7.8 dBm and a total radiated power of 0.9 mW is presented as the first demonstration of an integrated radiator with DPC capability. This prototype can control the polarization angle across the entire tuning range of 0 to 180 degrees while maintaining axial ratios above 10 dB, and control the axial ratio from 2.4 dB (near circular) to 14 dB (linear). We also demonstrate how simultaneous two-dimensional beam steering and DPC capabilities can even match the polarization to a mobile receiver antenna through a prototype 123-GHz 2x2 integrated DPC radiator array with a maximum EIRP of +12.3 dBm, polarization angle control across the full range of 0to 180 degrees as well as tunable axial ratio down to 1.2 dB and beam steering of up to 15 degrees in both dimensions. We also use slot-based DPC antennas to fabricate a 120-GHz integrated slot-based DPC radiator array, expected to have a maximum EIRP of +15.5 dBm.
We also introduce a new modulation scheme called Polarization Modulation (Pol-M) as a result of DPC capability, where the polarization itself is used for encoding the data. Pol-M is a spatial modulation method and is orthogonal to the existing phase and amplitude modulation schemes. Thus, it could be added on top of those schemes to enable creation of 4-D data constellations, or it can be used as the only basis for modulation to increase the stream security by misleading the undesired receivers. We discuss how DPC antenna enables Pol-M and also present PCB prototypes for Pol-M transmitter and receiver units operating at 2.4 GHz.</p
Advancing a Machine's Visual Awareness of People
Get PDFMethods to advance a machine's visual awareness of people with a focus on understanding 'who is where' in video are presented. 'Who' is used in a broad sense that includes not only the identity of a person but attributes of that person as well. Efforts are focused on improving algorithms in four areas of visual recognition: detection, tracking, fine-grained classification and person reidentification.
Each of these problems appear to be quite different on the surface; however, there are two broader questions that are answered across each of the works. The first, the machine is able to make better predictions when it has access to the extra information that is available in video. The second, that it is possible to learn on-the-fly from single examples. How each work contributes to answering these over-arching questions as well as its specific contributions to the relevant problem domain are as follows:
The first problem studied is one-shot, real-time, instance detection. Given a single image of a person, the task for the machine is to learn a detector that is specific to that individual rather than to an entire category such as faces or pedestrians. In subsequent images, the individual detector indicates the size and location of that particular person in the image. The learning must be done in real-time. To solve this problem, the proposed method starts with a pre-trained boosted category detector from which an individual-object detector is trained, with near-zero computational cost, through elementary manipulations of the thresholds of the category detector. Experiments on two challenging pedestrian and face datasets indicate that it is indeed possible to learn identity classifiers in real-time; besides being faster-trained, the proposed classifier has better detection rates than previous methods.
The second problem studied is real-time tracking. Given the initial location of a target person, the task for the machine is to determine the size and location of the target person in subsequent video frames, in real-time. The method proposed for solving this problem treats tracking as a repeated detection problem where potential targets are identified with a pre-trained boosted person detector and identity across frames is established by individual-specific detectors. The individual-specific detectors are learnt using the method proposed to solve the first problem. The proposed algorithm runs in real-time and is robust to drift. The tracking algorithm is benchmarked against nine state-of-the-art trackers on two benchmark datasets. Results show that the proposed method is 10% more accurate and nearly as fast as the fastest of the competing algorithms, and it is as accurate but 20 times faster than the most accurate of the competing algorithms.
The third problem studied is the fine-grained classification of people. Given an image of a person, the task for the machine is to estimate characteristics of that person such as age, clothing style, sex, occupation, social status, ethnicity, emotional state and/or body type. Since fine-grained classification using the entire human body is a relatively unexplored area, a large video dataset was collected. To solve this problem, a method that uses deep neural networks and video of a person is proposed. Results show that the class average accuracy when combining information from a sequence of images of an individual and then predicting the label is 3.5-7.1% better than independently predicting the label of each image, when severely under-represented classes are ignored.
The final problem studied is person reidentification. Given an image of a person, the task for the machine is to find images that match the identity of that person from a large set of candidate images. This is a challenging task since images of the same individual can vary significantly due to changes in clothing, viewpoint, pose, lighting and background. The method proposed for solving this problem is a two-stage deep neural network architecture that uses body part patches as inputs rather than an entire image of a person. Experiments show that rank-1 matching rates increase by 22-25.6% on benchmark datasets when compared to state-of-the-art methods.</p
Physiological and Biochemical Mechanisms of Phenazine-Mediated Survival in Pseudomonas aeruginosa
Get PDFThe opportunistic pathogen Pseudomonas aeruginosa secretes a class of colorful redox-active small molecules known as phenazines. Numerous functions have been proposed for phenazines, including antibiotic activity, virulence, cell-to-cell signaling, iron acquisition, and survival. This thesis delves into mechanisms of the latter role, that of long-term survival under oxidant-limiting conditions. Using a diverse array of methods, I investigated how phenazines support survival and how cells transfer electrons to phenazines, as well as the downstream effects that phenazines have on P. aeruginosa.
Direct measurements of NAD(H), ATP, the membrane potential, and fermentation products revealed that phenazines promote redox homeostasis and subsequently ATP synthesis. The ATP is used to maintain a membrane potential through the reverse action of the ATP synthase complex. Even though P. aeruginosa does not ferment on sugars, phenazines enable the anaerobic oxidation of glucose to acetate, suggesting P. aeruginosa may have previously under-appreciated metabolic flexibility in the absence of terminal electron acceptors. Activity assays with proteins purified natively from P. aeruginosa showed that glucose oxidation might be enabled in vivo by the pyruvate dehydrogenase complex, which can directly reduce phenazines using pyruvate as an electron donor. Liquid chromatography and mass spectrometry of culture supernatants showed that phenazines alter the chain length distribution of secreted quinolones, which may have indirect downstream signaling effects. Based on this result, combined with data from survival experiments, I hypothesize that phenazine-mediated redox homeostasis promotes β-oxidation and that fatty acid metabolism contributes to long-term survival. Further analysis also showed that P. aeruginosa cultures contain several previously-unreported sulfonated phenazines. In its natural environment, P. aeruginosa undoubtedly encounters other microbial species that consume or modify its phenazines. At least one of these, a Mycobacterium, contains a pyocyanin demethylating enzyme. The X-ray crystal structure of this protein revealed a novel reaction mechanism wherein the substrate is its own electron acceptor. Together, this work illuminates some of the many ways phenazines shape microbial communities in both clinical and environmental contexts.</p
Computational Modeling of the Mechanics of Elastic Structural Lattices: Effects of Lattice Architecture and Hierarchy
Get PDFThis thesis establishes advanced theoretical-computational techniques to understand and predict the mechanical properties of structural lattice metamaterials with a focus on the effective elastic properties. First, attention is devoted to the effective stiffness of hierarchical nanolattices, which depends on lattice topology, architecture, and inherent geometric imperfections. A computational substructuring technique is applied to predict the mechanics of hierarchical truss networks containing thousands to millions of truss members, with each solid, hollow-tube, or composite truss member requiring full-detail 3D resolution. By applying this methodology to hierarchical nanolattices structural hierarchy is shown to span several decades of relative density and effective stiffness with near-ideal effective stiffness scaling. Comparisons between experimental data and model predictions show convincing agreement and highlight the lattice sensitivity to fabrication-induced geometric imperfection. Second, elastic stress wave propagation in structural lattices is investigated with a focus on wave beaming (i.e., directional energy flow) under harmonic mechanical excitation. A new technique is introduced to obtain pseudo-continuous maps of group velocity magnitude vs. propagation direction vs. frequency to predict directional wave propagation, demonstrating traditional beaming prediction techniques are insufficient for many scenarios. The method is applied to two-dimensional structural lattices to predict directional energy flow. Predictions are verified by comparison to explicit dynamic simulations showing the limitations of the classical dispersion relation method. Overall, improved computational techniques are presented to better described, understand, predict and optimize the elastic behavior of truss lattices
Phase Boundary Mapping for Exploring New Thermoelectric Zintl Compounds
Get PDFUnderstanding and controlling the defect chemistry of bulk materials can vastly increase the opportunities for discovering highly efficient thermoelectrics. Good thermoelectrics are degenerate semiconductors and there are two types: n-type thermoelectrics, whose charge carriers are electrons, and p-type thermoelectrics conducting holes as carriers. Although normally one type can attain superior thermoelectric properties to the other depending on the electronic band structure of a material, a formation of the unfavorable defects sometimes prevents a material from obtaining the desired type. Similarly, even if the desired carrier type is realized, the Fermi level, which is a measure of the carrier density, could be kept from the optimum due to the formation of compensating defects. It has been known from growing binary semiconductors for electronics and optoelectronics such as GaAs and GaN that a growth condition can substantially alter the defect concentration of resulting samples. This is primarily due to the change in the reference atomic chemical potentials, but such defect engineering has not been utilized for the bulk thermoelectric research, resulting in overlooking the promising candidate materials.
In this work, we established an experimental methodology to fully explore all the accessible variations in chemical potentials of a target phase and demonstrated its implementation. Although a pursuance of purity of samples to be measured is a common experimental concept in solid-state chemistry, in practice, a small single-phase region of semiconductors allows samples to have a certain amount of impurities. Since different multi-phase equilibria have discrete chemical potentials, only when all the boundaries of the multi-phase equilibria around target phase are mapped out in nominal composition space are the measured transport properties of resulting samples properly correlated with their atomic chemical potential.
Utilizing this experimental concept we call "phase boundary mapping", we have identified the mechanism of obtaining the superior n-type conduction in Mg3Sb2-based compounds. To achieve their exceptionally high thermoelectric figure-of-merit (zT = 1.5 at 750 K), the formation energy of Mg-vacancy needs to be suppressed with excess Mg but this condition had been missing for over 80 years due to the absence of experimental concept to fully investigate properties of a material. Implementing phase boundary mapping has also allowed an inexpensive thermoelectric Zintl compound Ca9Zn4+xSb9 to be one of the best thermoelectrics in the intermediate temperature range (zT = 1.1). We have also successfully reduced the carrier concentration of Yb9Zn4+xSb9, which was originally thought to be impossible, leading to zT increased by a factor of five.</p
Nonlinear and Ultrafast Optical Investigations of Correlated Materials
Get PDFThis thesis comprises studies of 3d-5d transition metal oxides with various degrees of electronic correlation using nonlinear harmonic generation rotational anisotropy as well as time-resolved optical reflectivity methods. Specifically, we explored photo-induced phase transition in Ca2RuO4 and Sr2IrO4, discovered novel electronic phases in doped Sr2IrO4 and Sr3Ir2O7, and investigated different types of antiferromagnetic orders in transition metal trichalcogenides MPX3.</p
Chaucer the Beginner: Imagining Chaucer's Creative Process in Creating "Book of the Duchess"
Get PDFChaucer's first major poem, Book of the Duchess, represents a devoted reader's best attempt at creating a meaningful and authoritative text of their own. After extensive reading, Chaucer's first attempt to write a text draws significant inspiration from past texts such as Ovid's Metamorphoses and the medieval French dream vision Romance of the Rose. It is aspects of texts such as these which Chaucer mimics in order to imbue his own text with meaning and authority of its own. Chaucer's process of composing this first authoritative text will be discussed at length below, with a strong emphasis on the function of voice and imitation, and how inspiration, a simultaneous source of ideas and motivation, drives Chaucer's entire creative process. Chaucer utilizes and manipulates a great many voices in Book of the Duchess, many of which obviously don't belong to Chaucer but rather to authors of the past. It is these multiple already-authoritative voices, in addition to a single nameless voice narrating the story, that ultimately constitute Chaucer's own, original voice.</p
Single Lens Holographic Imaging
No full textMetasurface phase masks offer a new way of imaging in that the lens sizes are orders of magnitude smaller and allow for batch fabrication of a large number of miniature lenses. Unfortunately, current metasurface lens technology cannot simultaneously overcome the issues of image distortion for objects far from the optical axis and chromatic aberrations. Using simulations in Matlab, we study and design a single lens holographic system using metasurface lenses to drastically reduce both off-axis distortions and chromatic aberrations. Using a single metasurface lens, we demonstrate that a high quality image can be obtained for an object with a 30° angular diameter and 123 fractional bandwidth. We also experimentally demonstrate reduction of aberrations and 3D imaging for spare images as well as show mediocre quality reconstructed dense images
Symmetry and Variational Analyses of Fluid Interface Equations in the Thin Film Limit
Get PDFThis thesis concerns a class of nonlinear partial differential equations up to fourth order in spatial derivatives that models thin viscous films. In Chapter 1, we review the derivations of thin film equations from the fundamental transport equations. Section 1.1 contains the derivation for a thermocapillary driven film to familiarize the reader with the key long-wavelength approximation that has been successful in modeling a myriad of thin viscous films. In Section 1.2, we consider the coupling between a thin viscous layer and a much thicker fluid layer with much larger viscosity and conductivity and show how a novel, non-local thermocapillary thin film equation can be derived to model such a system. We then review the wider class of thin film equations in Section 1.3, note the important Cahn-Hilliard variational form of these equations, and demonstrate that classic mathematical results concerning the inverse problem of the calculus of variations permit an algorithmic procedure for discovering Lyapunov functionals. In Chapter 2, we review applications of symmetry methods to partial differential equations. Section 2.1 contains an original geometrical motivation for the study of self-similar reductions which draws an analogy with the fixed points of dynamical systems. In Section 2.2, we derive for the first time the full set of symmetries of the fully two-dimensional thin film equations. We then enumerate the possible symmetry reductions of the thin film equations, and discover several which have not been previously recognized. In Chapter 3, we consider rotationally invariant, steady droplet solutions and their stability. In Section 3.1, we derive stability criteria for thermocapillary-driven droplets, and show a novel correspondence between droplet stability, droplet volume, and droplet Lyapunov energy. We consider thin films under other forces in Section 3.2 and make new predictions about conditions under which such films develop into droplets, columns, or jets of fluid. In Chapter 4, we consider the scale invariant symmetry reductions of thin film equations. In Section 4.1 we describe the extraordinarily rich variety of such solutions in the spreading of a insoluble surfactant on a thin viscous film, identify previously unrecognized scale invariant solutions which are well-behaved at the origin, and demonstrate their relevance with finite element simulations. Lastly, in Section 4.2, we illustrate for capillary driven films some numerical solutions to the novel reductions we uncovered in Chapter 2. Each chapter concludes with a Notes section which summarizes the new results contained therein and relates them to the wider literature