DSpace at MIT
This site is a university repository providing access to the publication output of the institution. Registered users can set up email alerts to notify them of newly added relevant content. A certain level of encryption and security is embedded in the site which may cause some users accessibility problems.
Electrical Engineering and Computer Sciences - Master's degree
Mostrando recursos 1 - 20 de 4.590
Interactive visualization of big data leveraging databases for scalable computation - Battle, Leilani Marie
Modern database management systems (DBMS) have been designed to efficiently store, manage and perform computations on massive amounts of data. In contrast, many existing visualization systems do not scale seamlessly from small data sets to enormous ones. We have designed a three-tiered visualization system called ScalaR to deal with this issue. ScalaR dynamically performs resolution reduction when the expected result of a DBMS query is too large to be effectively rendered on existing screen real estate. Instead of running the original query, ScalaR inserts aggregation, sampling or filtering operations to reduce the size of the result. This thesis presents the...
Estimating the material properties of fabric through the observation of motion - Bouman, Katherine L. (Katherine Louise)
We present a framework for predicting the physical properties of moving deformable objects observed in video. We apply our framework to analyze videos of fabrics moving under various unknown wind forces, and recover two key material properties of the fabric: stiffness and mass. We extend features previously developed to compactly represent static image textures to describe video textures such as fabric motion. A discriminatively trained regression model is then used to predict the physical properties of fabric from these features. The success of our model is demonstrated on a new database of fabric videos with corresponding measured ground truth material...
Aerial reconstructions via probabilistic data fusion - Cabezas, Randi
In this thesis we propose a probabilistic model that incorporates multi-modal noisy measurements: aerial images and Light Detection and Ranging (LiDAR) to recover scene geometry and appearance in order to build a 3D photo-realistic model of a given scene. In urban environments, these reconstructions have many applications, such as surveillance, and urban planning. The proposed probabilistic model can be viewed as a data fusion model, in which the two data sources complement each other and allow for better results than when only a single one is present. Moreover, this modeling approach has the advantages that it can capture uncertainty in...
An OpenGL backend for Halide - Chornay, Nicholas J
High performance image processing requires not only an efficient underlying algorithm but also an implementation tailored to maximally exploit the available hardware resources. In practice, this requires low-level optimization, platform-specific instructions, and, when available, the use of special purpose hardware such as GPU. Halide is a domain-specific programming language targeted at image processing applications. Its programming model decouples an algorithm from the details of its execution, vastly simplifying development and optimization. We present an OpenGL backend for the Halide compiler, which enables Halide programs to run GPU computation on devices that support the OpenGL API. In particular, this paves the...
A 19-bit monolithic charge-balancing A/D converter - Chew, Tung Shen
Existing high-resolution ADC topologies are expensive, complicated and vulnerable to switch leakage at high temperature. This thesis introduces the Modified Landsburg ADC, a high-resolution converter optimized for minimum cost and die area. Switch resistance cancellation, charge-injection compensation and auto-zero methods are used to build a simple and robust ADC which will operate reliably across the military temperature range. Implemented on a 0.25[mu]m CMOS process, with a die area of under 300mil², the Modified Landsburg ADC achieves an ENOB of 15.6 bits and an SNR of 95.8dB at 5 Samples/s while consuming <2mW. It enables the integration of a high-resolution converter...
Integrated DC-DC converter with ultra-low quiescent current - Chen, Christine, M. Eng. Massachusetts Institute of Technology
Based on the LTC3588, the design of a bandgap reference and a comparator for use in the control circuitry of DC-DC converter with an ultra-low quiescent current of 150nA is presented here. Not only will this thesis discuss the challenges encountered over the course of designing circuits to operate at such low current levels, but it will also provide proof of concept silicon evaluation data of modified LTC3588 chips demonstrating that such low current operation is viable.
Word sense disambiguation in clinical text - Chasin, Rachel (Rachel G.)
Lexical ambiguity, the ambiguity arising from a string with multiple meanings, is pervasive in language of all domains. Word sense disambiguation (WSD) and word sense induction (WSI) are the tasks of resolving this ambiguity. Applications in the clinical and biomedical domain focus on the potential disambiguation has for information extraction. Most approaches to the problem are unsupervised or semi-supervised because of the high cost of obtaining enough annotated data for supervised learning. In this thesis we compare the application of a semi-supervised general domain state of the art WSI method to clinical text to the best known knowledge-based unsupervised methods...
Efficient, low-distortion switch-mode power amplifier for amplitude modulation - Buenrostro, Christopher (Christopher B.)
Amplitude modulation systems are used in many areas of engineering, especially communication based disciplines. Typical systems have low-level signals which encode the desired information to be transmitted. These low-level signals are usually not of adequate power to transmit the desired information across a target medium, thus requiring modulation and power amplification. In general, two goals of power amplifier design are low-distortion and high-efficiency. To meet these goals, a unique switch-mode power amplifier intended for amplitude modulation which utilizes spectral content at the carrier frequency is designed, simulated, and built. Theoretical predictions of total harmonic distortion (THD) and efficiency are made,...
Spec2Fab : a reducer-tuner model for translating specifications to 3D prints - Chen, Desai
Multi-material 3D printing allows objects to be composed of complex, heterogeneous arrangements of materials. It is often more natural to define a functional goal than to define the material composition of an object. Translating these functional requirements to fabricable 3D prints is still an open research problem. Recently, several specific instances of this problem have been explored (e.g., appearance or elastic deformation), but they exist as isolated, monolithic algorithms. In this research, I propose an abstraction mechanism that simplifies the design, development, implementation, and reuse of these algorithms. The solution relies on two new data structures: a reducer tree that...
Trading isolation for certifiable randomness expansion - Coudron, Matthew Ryan
A source of random bits is an important resource in modern cryptography, algorithms and statistics. Can one ever be sure that a "random" source is truly random, or in the case of cryptography, secure against potential adversaries or eavesdroppers? Recently the study of non-local properties of entanglement has produced an interesting new perspective on this question, which we will refer to broadly as Certifiable Randomness Expansion (CRE). CRE refers generally to a process by which a source of information-theoretically certified randomness can be constructed based only on two simple assumptions: the prior existence of a short random seed and the...
Sensitive, 3D micromotion compensation in a surface-electrode ion trap - Eltony, Amira M. (Amira Madeleine)
Following successful demonstrations of quantum algorithms and error correction with a handful of trapped ions in a macroscopic, machined Paul trap, there is a growing effort to move towards microfabricated traps with all the electrodes on a single chip. These traps, known as surface-electrode ion traps, are more amenable to being shrunk in size and replicated, or integrated with optical components and electronic devices. However, in the shift towards surface-electrode traps, and as traps are miniaturized in general, laser beams are brought closer to electrode surfaces, exacerbating laser-induced charging. Because of their charge, trapped ions are extremely sensitive to stray...
Bayesian distance metric learning on i-vector for speaker verification - Fang, Xiao, S.M. Massachusetts Institute of Technology
This thesis explores the use of Bayesian distance metric learning (Bayes_dml) for the task of speaker verification using the i-vector feature representation. We propose a framework that explores the distance constraints between i-vector pairs from the same speaker and different speakers. With an approximation of the distance metric as a weighted covariance matrix of the top eigenvectors from the data covariance matrix, variational inference is used to estimate a posterior distribution of the distance metric. Given speaker labels, we select different-speaker data pairs with the highest cosine scores to form a different-speaker constraint set. This set captures the most discriminative...
Dimension reduction algorithms for near-optimal low-dimensional embeddings and compressive sensing - Grant, Elyot
In this thesis, we establish theoretical guarantees for several dimension reduction algorithms developed for applications in compressive sensing and signal processing. In each instance, the input is a point or set of points in d-dimensional Euclidean space, and the goal is to find a linear function from Rd into Rk , where k << d, such that the resulting embedding of the input pointset into k-dimensional Euclidean space has various desirable properties. We focus on two classes of theoretical results: -- First, we examine linear embeddings of arbitrary pointsets with the aim of minimizing distortion. We present an exhaustive-search-based algorithm...
Transmission scheduling with deadline and throughput constraints - Kim, Kyu Seob
Today wireless networks are increasingly used to perform applications with Quality of Service constraints (QoS) such as delay, delivery ratio, and channel reliability. Especially, as demand for real-time transmissions increases, transmitting packets with hard delivery deadlines poses an important network control problem. In this thesis, we propose a framework for characterizing feasibility regions and finding an optimal scheduling policy in various wireless networks. First, we start with a wireless network with multiple unicast flows. We investigate how delay in feedback information decreases the feasibility region. Second, we consider time-varying channels and how delay in network state information decreases the feasibility...
Distributed optimization of traffic delay on a periodic switched grid network - Kozynski Waserman, Fabián Ariel
Given a switched network, e.g. a city grid with semaphores in its intersections or a packet network, each unit (car or packet) accumulates some delay while traversing the network. This delay is undesirable but unavoidable, which makes minimizing the total average delay of the network given certain constraints, a desirable objective. In this work, we consider the case of periodic networks, meaning that in every traffic cycle inputs to the system are the same and we try to arrive to an allocation of phases in every intersection that minimizes the total delay per cycle. We propose a model for such...
Execution model and optimizing compilation for execution migration - Lebedev, Ilia Andreevich
Although systems with hardware support for fine-grained execution migration are becoming a reality, no concrete execution model or compiler exist for these machines. This limits the complexity of software that can be written for these machines, and therefore also the scope of studies for which these machines can be used. In this thesis, we define a productive programming model for an execution migration platform by exposing migration as a set of interfaces usable with the C programming language via a custom optimizing compiler. We employ hardware-software co-design to describe a stack core architecture with support for partial context migration in...
QT-interval adaptation to changes in autonomic balance - Nosakhare, Ehimwenma
ECG variability, as it relates to the influence of the autonomic nervous system on the heart, is primarily studied via frequency-domain and time-domain analysis of heart rate variability (HRV). HRV studies the variability of the RR intervals in the ECG; these intervals are modulated by the autonomic influence on the periodicity of the the heart's pacemaker, the sino-atrial node. The autonomic influence at this level is dominated by the parasympathetic nervous system. In order to have a robust assessment of autonomic balance, there is a need for an ECG-based approach to assess the influence of the sympathetic nervous system. In...
Replacing cellular with WiFi direct communication for a highly interactive, high bandwidth multiplayer game - Ortiz, Pablo (Pablo Jose)
The objective of this work is to explore the benefits of replacing cellular with Wi-Fi Direct communication in mobile applications. Cellular connections consume significant power on mobile devices and are too slow for many highly interactive mobile applications. Wi-Fi Direct, the recently released wireless standard, promises to provide the speed, power efficiency, and security of Wi-Fi to devices communicating within a short range. Using Wi-Fi performance as a baseline, the performance of a proof of concept system, Super Tux Kart Direct, is evaluated when communication is enabled by either LTE or Wi-Fi Direct. At the time of writing and to...
An analytical framework for field electron emission, incorporating quantum- confinement effects - Patterson. Alex A. (Alex Andrew)
As field electron emitters shrink to nanoscale dimensions, the effects of quantum confinement of the electron supply and electric field enhancement at the emitter tip play a significant role in determining the emitted current density (ECD). Consequently, the Fowler-Nordheim (FN) equation, which primarily applies to field emission from the planar surface of a bulk metal may not be valid for nanoscale emitters. While much effort has focused on studying emitter tip electrostatics, not much attention has been paid to the consequences of a quantum-confined electron supply. This work builds an analytical framework from which ECD equations for quantum-confined emitters of...
Design of class E resonant rectifiers for very high frequency power conversion - Santiago-GonzáIez, Juan Antonio
Resonant rectifiers have important applications in very-high-frequency power conversion systems, including dc-dc converters, wireless power transfer systems, and energy recovery circuits for radio-frequency systems. In many of these applications, it is desirable for the rectifier to appear as a resistor at its ac input port. However, for a given dc output voltage, the input impedance of a resonant rectifier varies in magnitude and phase as output power changes. In this thesis, a design method is introduced for realizing single-diode "shunt-loaded" resonant rectifiers, or class E rectifiers, that provide near-resistive input impedance over a wide range of output power levels. The...