UMD Theses and Dissertations
Permanent URI for this collectionhttp://hdl.handle.net/1903/3
New submissions to the thesis/dissertation collections are added automatically as they are received from the Graduate School. Currently, the Graduate School deposits all theses and dissertations from a given semester after the official graduation date. This means that there may be up to a 4 month delay in the appearance of a given thesis/dissertation in DRUM.
More information is available at Theses and Dissertations at University of Maryland Libraries.
Browse
6 results
Search Results
Item Outdoor Localization and Path Planning for Repositioning a Self-Driving Electric Scooter(2023) Poojari, Srijal Shekhar; Paley, Derek; Systems Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)The long-term goal of this research is to develop self-driving e-scooter technology to increase sustainability, accessibility, and equity in urban mobility. Toward this goal, in this work, we design and implement a self-driving e-scooter with the ability to safely travel along a pre-planned route using automated, onboard control without a rider. We also construct an autonomous driving framework by synthesizing open-source robotics software libraries with custom-designed modules specific to an e-scooter, including path planning and state estimation. The hardware and software development steps along with design choices and pitfalls are documented. Results of real-world autonomous navigation of our retrofitted e-scooter, along with the effectiveness of our localization methods are presented.Item Distributed Passive Sensor Network for the Geolocation of RF Emitters(2019) Dillon, Matthew; Ephremides, Anthony; Electrical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)The ability to localize an RF emitter has emerged in both commercial and military technology, and is an important capability in modern cognitive radios to achieve spectral awareness. Of importance, is the accuracy of the geolocation of the RF emitter. In this thesis, we address the blind localization problem given a network of software-defined radio receivers that monitor the spectrum to determine the presence of an unknown emitter. We discuss the underlying challenges and various approaches to the geolocation problem that can be utilized. In particular, two algorithms that are used extensively in literature are investigated: time-difference of arrival, and power-difference of arrival. In the first part of the thesis, the algorithms are presented, and the error performance is characterized analytically, and then conducted through simulation. A more robust method which implements the fusion of both algorithms for an improved estimation. In the second part, we conduct a small- scale laboratory emulation of the geolocation algorithms on a network of radios to contrast the simulation results of the algorithms from the emulation results. The results provided insight to the shortcomings of each algorithm, and potential extensions for further accuracy improvement.Item Computational and Analytical Investigations of Disordered and Interacting Systems(2013) Biddle, John Charles; Das Sarma, Sankar; Physics; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Localization of particle wavefunctions in quasi-disordered one dimensional incommensurate lattices is studied both numerically and analytically. Through exact diagonalization, we show that energy dependent mobility edges can appear in the case of shallow lattices. We also show that these mobility edges can be studied with a tight-binding model (an extension of the Aubry-Andre model) that has energy dependent mobility edges that can be determined analytically. Topological aspects of the Aubry-Andre/Harper model are also studied by numerically calculating the Chern number. We first verify arguments by numerical calculations that variations in the Chern density decrease with increasing system size when the potential is incommensurate with the lattice. Next we introduce random disorder into the model and study the Chern number and the Chern density as a function of disorder strength by using the non-commutative Brillouin zone. We show that variations of the Chern density take on the same trends for both commensurate and incommensurate potentials after some critical disorder strength is reached. Strongly correlated quantum Hall states are also examined. We numerically examine the entanglement entropy and the entanglement spectrum of fractional quantum hall states as a function of the finite layer thickness $d$ of the quasi-two-dimensional system for a number of filling fractions $\nu$ in the lowest and the second Landau levels: $\nu$ = 1/3, 7/3, 1/2, and 5/2. We observe that the entanglement measures are dependent on which Landau level the electrons fractionally occupy and are completely consistent with the results based on wavefunction overlap calculations. We also compare the ground state energies by variational Monte Carlo of the spin unpolarized Halperin 331 and the spin polarized Moore-Read (MR) Pfaffian fractional quantum Hall states at half filling of the lowest Landau level (LLL) and the second Landau level (SLL) as a function of small deviations around the Coulomb point. Our results suggest that even under moderate deviations in the interaction potential the MR Pfaffian description is more energetically favorable than the Halperin 331 state in the half filled SLL, consistent with recent experimental investigations.Item CHARGE TRANSPORT IN GRAPHENE WITH ADATOM OVER-LAYERS ; CHARGED IMPURITY SCATTERING, DIELECTRIC SCREENING, AND LOCALIZATION.(2011) Jang, Chaun; Fuhrer, Michael S; Physics; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Graphene, a single atom thick plane of graphite, is a novel two-dimensional electron system in which the low-energy electrons behave as massless chiral Dirac fermions. This thesis explores the effects of disorder in graphene through controlled surface modification in ultra-high vacuum (UHV), coupled with in situ electronic transport experiments. Three different roles of adatom overlayers on graphene are investigated. First, the effects of charged impurity scattering are studied by introducing potassium ions on the graphene at low temperature in UHV. The theoretically expected magnitude and linear density-dependence of the conductivity due to long range Coulomb scattering is verified. Second, the effective dielectric constant of graphene is modified by adding ice overlayers at low temperature in UHV. The opposing effects of screening on scattering by long range (charged impurity) and short range impurities are observed as variations in conductivity, and the changes are in agreement with Boltzmann theory for graphene transport within the random phase approximation. The minimum conductivity of graphene is roughly independent of charged impurity density and dielectric constant, in agreement with the self-consistent theory of screened carrier density inhomogeneity (electron and hole puddles). Taken together, the experimental results on charged impurity scattering and dielectric screening strongly support that long range Coulomb scattering is the dominant scattering mechanism in as-fabricated graphene on SiO2. In addition to the semi-classical transport properties, quantum transport is also studied with cobalt decorated graphene. Strong localization is achieved in the disordered graphene through deposition of cobalt nanoclusters. In finite magnetic field a phase transition occurs from the localized state to the quantum Hall state. Scaling analysis confirms that the transition is a quantum phase transition which is similar to the localization - delocalization transitions in other two dimensional electron systems.Item Estimation of Elevation and Azimuth in a Neuromorphic VLSI Bat Echolocation System(2009) Abdalla, Hisham Ahmed Nabil; Horiuchi, Timothy K; Electrical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Auditory localization is an interesting and challenging problem; the location of the sound source is not spatially encoded in the peripheral sensory system as it is in the visual or somatosensory systems. Instead it must be computed from the neural representation of the sound reaching both ears. Echolocation is a form of auditory localization, however, an important distinction is that the sound being localized is an echo of the sound emitted by the animal itself. This dissertation presents a neuromorphic VLSI circuit model of a bat echolocation system. The acoustic cues that we use in our system are the binaural interaural level differences (ILDs) and the monaural spectral cues. We have designed an artificial bat head using 3D CAD software and fabricated it using a 3D printer. The artificial bat head is capable of generating the necessary acoustic cues for localization. We have designed and fabricated an ultrasonic cochlea chip with 16 cochlear filters and 128 spiking cochlear neurons (eight neurons per cochlear filter), the cochlear filters and neurons transform the analog input into a spike-based cochlear representation. We have also designed and fabricated two feature extraction chips: a monaural spectral difference chip and a binaural ILD chip, that together can extract the localization cues from the spike-based cochlear representation. The monaural spectral difference chip consists of 240 spiking neurons; each neuron compares the activity of two cochlear filters within the same ear. The binaural ILD chip consists of 32 spiking neurons (two per cochlear filter) that model the processing that takes place in the lateral superior olive (LSO). We demonstrate that the spatiotemporal pattern of spiking outputs from the feature extraction chips can be decoded to estimate the direction (elevation and azimuth) of an ultrasonic chirp.Item SALAM: A SCALABLE ANCHOR-FREE LOCALIZATION ALGORITHM FOR WIRELESS SENSOR NETWORKS(2006-04-26) Youssef, Adel Amin Abdel Azim; Agrawala, Ashok K; Computer Science; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)In this dissertation, we present SALAM, a scalable anchor-free protocol for localization in wireless sensor networks. SALAM can determine the positions of sensor nodes without any infrastructure support. We assume that each node has the capability to estimate distances to its corresponding neighbors, that are within its transmission range. SALAM allows to trade the accuracy of the estimated position against node transmission range and/or computational power. The application layer can choose from a whole range of different options, to estimate the sensor node's positions with different accuracy while conserving battery power. Scalability is achieved by dividing the network into overlapping multi-hop clusters each with its own cluster head node. Each cluster head is responsible for building a local relative map corresponding to its cluster using intra-cluster node's range measurements. To obtain the global relative topology of the network, the cluster head nodes collaboratively combine their local maps using simple matrix transformations. In order for two cluster heads to perform a matrix transformation, there must be at least three boundary nodes that belongs to both clusters (i.e. the two clusters are overlapping with degree 3). We formulate the overlapping multi-hop clustering problem and present a randomized distributed heuristic algorithm for solving the problem. We evaluate the performance of the proposed algorithm through analytical analysis and simulation. A major problem with multi-hop relative location estimation is the error accumulated in the node position as it becomes multi-hop away from the cluster head node. We analyze different sources of error and develop techniques to avoid these errors. We also show how the local coordinate system (LCS) affects the accuracy and propose different heuristics to select the LCS. Simulation results show that SALAM achieves precise localization of sensors. We show that our approach is scalable in terms of communication overhead regardless of the network size. In addition, we capture the impact of different parameters on the accuracy of the estimated node's positions. The results also show that SALAM is able to achieve accuracy better than the current ad-hoc localization algorithms.