Matching Items (119)
Description
Effective modeling of high dimensional data is crucial in information processing and machine learning. Classical subspace methods have been very effective in such applications. However, over the past few decades, there has been considerable research towards the development of new modeling paradigms that go beyond subspace methods. This dissertation focuses on the study of sparse models and their interplay with modern machine learning techniques such as manifold, ensemble and graph-based methods, along with their applications in image analysis and recovery. By considering graph relations between data samples while learning sparse models, graph-embedded codes can be obtained for use in unsupervised, supervised and semi-supervised problems. Using experiments on standard datasets, it is demonstrated that the codes obtained from the proposed methods outperform several baseline algorithms. In order to facilitate sparse learning with large scale data, the paradigm of ensemble sparse coding is proposed, and different strategies for constructing weak base models are developed. Experiments with image recovery and clustering demonstrate that these ensemble models perform better when compared to conventional sparse coding frameworks. When examples from the data manifold are available, manifold constraints can be incorporated with sparse models and two approaches are proposed to combine sparse coding with manifold projection. The improved performance of the proposed techniques in comparison to sparse coding approaches is demonstrated using several image recovery experiments. In addition to these approaches, it might be required in some applications to combine multiple sparse models with different regularizations. In particular, combining an unconstrained sparse model with non-negative sparse coding is important in image analysis, and it poses several algorithmic and theoretical challenges. A convex and an efficient greedy algorithm for recovering combined representations are proposed. Theoretical guarantees on sparsity thresholds for exact recovery using these algorithms are derived and recovery performance is also demonstrated using simulations on synthetic data. Finally, the problem of non-linear compressive sensing, where the measurement process is carried out in feature space obtained using non-linear transformations, is considered. An optimized non-linear measurement system is proposed, and improvements in recovery performance are demonstrated in comparison to using random measurements as well as optimized linear measurements.
ContributorsNatesan Ramamurthy, Karthikeyan (Author) / Spanias, Andreas (Thesis advisor) / Tsakalis, Konstantinos (Committee member) / Karam, Lina (Committee member) / Turaga, Pavan (Committee member) / Arizona State University (Publisher)
Created2013
Description
Vector Fitting (VF) is a recent macromodeling method that has been popularized by its use in many commercial software for extracting equivalent circuit's of simulated networks. Specifically for material measurement applications, VF is shown to estimate either the permittivity or permeability of a multi-Debye material accurately, even when measured in the presence of noise and interferences caused by test setup imperfections. A brief history and survey of methods utilizing VF for material measurement will be introduced in this work. It is shown how VF is useful for macromodeling dielectric materials after being measured with standard transmission line and free-space methods. The sources of error in both an admittance tunnel test device and stripline resonant cavity test device are identified and VF is employed for correcting these errors. Full-wave simulations are performed to model the test setup imperfections and the sources of interference they cause are further verified in actual hardware measurements. An accurate macromodel is attained as long as the signal-to-interference-ratio (SIR) in the measurement is sufficiently high such that the Debye relaxations are observable in the data. Finally, VF is applied for macromodeling the time history of the total fields scattering from a perfectly conducting wedge. This effort is an initial test to see if a time domain theory of diffraction exists, and if the diffraction coefficients may be exactly modeled with VF. This section concludes how VF is not only useful for applications in material measurement, but for the solution of modeling fields and interactions in general.
ContributorsRichards, Evan (Author) / Diaz, Rodolfo E (Thesis advisor) / Tsakalis, Konstantinos (Committee member) / Platte, Rodrigo (Committee member) / Arizona State University (Publisher)
Created2013
Description
Advances in implantable MEMS technology has made possible adaptive micro-robotic implants that can track and record from single neurons in the brain. Development of autonomous neural interfaces opens up exciting possibilities of micro-robots performing standard electrophysiological techniques that would previously take researchers several hundred hours to train and achieve the desired skill level. It would result in more reliable and adaptive neural interfaces that could record optimal neural activity 24/7 with high fidelity signals, high yield and increased throughput. The main contribution here is validating adaptive strategies to overcome challenges in autonomous navigation of microelectrodes inside the brain. The following issues pose significant challenges as brain tissue is both functionally and structurally dynamic: a) time varying mechanical properties of the brain tissue-microelectrode interface due to the hyperelastic, viscoelastic nature of brain tissue b) non-stationarities in the neural signal caused by mechanical and physiological events in the interface and c) the lack of visual feedback of microelectrode position in brain tissue. A closed loop control algorithm is proposed here for autonomous navigation of microelectrodes in brain tissue while optimizing the signal-to-noise ratio of multi-unit neural recordings. The algorithm incorporates a quantitative understanding of constitutive mechanical properties of soft viscoelastic tissue like the brain and is guided by models that predict stresses developed in brain tissue during movement of the microelectrode. An optimal movement strategy is developed that achieves precise positioning of microelectrodes in the brain by minimizing the stresses developed in the surrounding tissue during navigation and maximizing the speed of movement. Results of testing the closed-loop control paradigm in short-term rodent experiments validated that it was possible to achieve a consistently high quality SNR throughout the duration of the experiment. At the systems level, new generation of MEMS actuators for movable microelectrode array are characterized and the MEMS device operation parameters are optimized for improved performance and reliability. Further, recommendations for packaging to minimize the form factor of the implant; design of device mounting and implantation techniques of MEMS microelectrode array to enhance the longevity of the implant are also included in a top-down approach to achieve a reliable brain interface.
ContributorsAnand, Sindhu (Author) / Muthuswamy, Jitendran (Thesis advisor) / Tillery, Stephen H (Committee member) / Buneo, Christopher (Committee member) / Abbas, James (Committee member) / Tsakalis, Konstantinos (Committee member) / Arizona State University (Publisher)
Created2013
Description
Mobile robots are used in a broad range of application areas; e.g. search and rescue, reconnaissance, exploration, etc. Given the increasing need for high performance mobile robots, the area has received attention by researchers. In this thesis, critical control and control-relevant design issues for differential drive mobile robots is addressed. Two major themes that have been explored are the use of kinematic models for control design and the use of decentralized proportional plus integral (PI) control. While these topics have received much attention, there still remain critical questions which have not been rigorously addressed. In this thesis, answers to the following critical questions are provided: When is 1. a kinematic model sufficient for control design? 2. coupled dynamics essential? 3. a decentralized PI inner loop velocity controller sufficient? 4. centralized multiple-input multiple-output (MIMO) control essential? and how can one design the robot to relax the requirements implied in 1 and 2? In this thesis, the following is shown: 1. The nonlinear kinematic model will suffice for control design when the inner velocity (dynamic) loop is much faster (10X) than the slower outer positioning loop. 2. A dynamic model is essential when the inner velocity (dynamic) loop is less than two times faster than the slower outer positioning loop. 3. A decentralized inner loop PI velocity controller will be sufficient for accomplish- ing high performance control when the required velocity bandwidth is small, rel- ative to the peak dynamic coupling frequency. A rule-of-thumb which depends on the robot aspect ratio is given. 4. A centralized MIMO velocity controller is needed when the required bandwidth is large, relative to the peak dynamic coupling frequency. Here, the analysis in the thesis is sparse making the topic an area for future analytical work. Despite this, it is clearly shown that a centralized MIMO inner loop controller can offer increased performance vis- ́a-vis a decentralized PI controller. 5. Finally, it is shown how the dynamic coupling depends on the robot aspect ratio and how the coupling can be significantly reduced. As such, this can be used to ease the requirements imposed by 2 and 4 above.
ContributorsAnvari, Iman (Author) / Rodriguez, Armando A (Thesis advisor) / Si, Jenni (Committee member) / Tsakalis, Konstantinos (Committee member) / Arizona State University (Publisher)
Created2013
Description
Approximately 1% of the world population suffers from epilepsy. Continuous long-term electroencephalographic (EEG) monitoring is the gold-standard for recording epileptic seizures and assisting in the diagnosis and treatment of patients with epilepsy. However, this process still requires that seizures are visually detected and marked by experienced and trained electroencephalographers. The motivation for the development of an automated seizure detection algorithm in this research was to assist physicians in such a laborious, time consuming and expensive task. Seizures in the EEG vary in duration (seconds to minutes), morphology and severity (clinical to subclinical, occurrence rate) within the same patient and across patients. The task of seizure detection is also made difficult due to the presence of movement and other recording artifacts. An early approach towards the development of automated seizure detection algorithms utilizing both EEG changes and clinical manifestations resulted to a sensitivity of 70-80% and 1 false detection per hour. Approaches based on artificial neural networks have improved the detection performance at the cost of algorithm's training. Measures of nonlinear dynamics, such as Lyapunov exponents, have been applied successfully to seizure prediction. Within the framework of this MS research, a seizure detection algorithm based on measures of linear and nonlinear dynamics, i.e., the adaptive short-term maximum Lyapunov exponent (ASTLmax) and the adaptive Teager energy (ATE) was developed and tested. The algorithm was tested on long-term (0.5-11.7 days) continuous EEG recordings from five patients (3 with intracranial and 2 with scalp EEG) and a total of 56 seizures, producing a mean sensitivity of 93% and mean specificity of 0.048 false positives per hour. The developed seizure detection algorithm is data-adaptive, training-free and patient-independent. It is expected that this algorithm will assist physicians in reducing the time spent on detecting seizures, lead to faster and more accurate diagnosis, better evaluation of treatment, and possibly to better treatments if it is incorporated on-line and real-time with advanced neuromodulation therapies for epilepsy.
ContributorsVenkataraman, Vinay (Author) / Jassemidis, Leonidas (Thesis advisor) / Spanias, Andreas (Thesis advisor) / Tsakalis, Konstantinos (Committee member) / Arizona State University (Publisher)
Created2012
Description
This report provides an overview of scramjet-powered hypersonic vehicle modeling and control challenges. Such vehicles are characterized by unstable non-minimum phase dynamics with significant coupling and low thrust margins. Recent trends in hypersonic vehicle research are summarized. To illustrate control relevant design issues and tradeoffs, a generic nonlinear 3DOF longitudinal dynamics model capturing aero-elastic-propulsive interactions for wedge-shaped vehicle is used. Limitations of the model are discussed and numerous modifications have been made to address control relevant needs. Two different baseline configurations are examined over a two-stage to orbit ascent trajectory. The report highlights how vehicle level-flight static (trim) and dynamic properties change over the trajectory. Thermal choking constraints are imposed on control system design as a direct consequence of having a finite FER margin. The implication of this state-dependent nonlinear FER margin constraint, the right half plane (RHP) zero, and lightly damped flexible modes, on control system bandwidth (BW) and FPA tracking has been discussed. A control methodology has been proposed that addresses the above dynamics while providing some robustness to modeling uncertainty. Vehicle closure (the ability to fly a trajectory segment subject to constraints) is provided through a proposed vehicle design methodology. The design method attempts to use open loop metrics whenever possible to design the vehicle. The design method is applied to a vehicle/control law closed loop nonlinear simulation for validation. The 3DOF longitudinal modeling results are validated against a newly released NASA 6DOF code.
ContributorsDickeson, Jeffrey James (Author) / Rodriguez, Armando A (Thesis advisor) / Tsakalis, Konstantinos (Committee member) / Si, Jennie (Committee member) / Wells, Valana (Committee member) / Kawski, Mattias (Committee member) / Arizona State University (Publisher)
Created2012
Description
Development of post-traumatic epilepsy (PTE) after traumatic brain injury (TBI) is a major health concern (5% - 50% of TBI cases). A significant problem in TBI management is the inability to predict which patients will develop PTE. Such prediction, followed by timely treatment, could be highly beneficial to TBI patients. Six male Sprague-Dawley rats were subjected to a controlled cortical impact (CCI). A 6mm piston was pneumatically driven 3mm into the right parietal cortex with velocity of 5.5m/s. The rats were subsequently implanted with 6 intracranial electroencephalographic (EEG) electrodes. Long-term (14-week) continuous EEG recordings were conducted. Using linear (coherence) and non-linear (Lyapunov exponents) measures of EEG dynamics in conjunction with measures of network connectivity, we studied the evolution over time of the functional connectivity between brain sites in order to identify early precursors of development of epilepsy. Four of the six TBI rats developed PTE 6 to 10 weeks after the initial insult to the brain. Analysis of the continuous EEG from these rats showed a gradual increase of the connectivity between critical brain sites in terms of their EEG dynamics, starting at least 2 weeks prior to their first spontaneous seizure. In contrast, for the rats that did not develop epilepsy, connectivity levels did not change, or decreased during the whole course of the experiment across pairs of brain sites. Consistent behavior of functional connectivity changes between brain sites and the "focus" (site of impact) over time was demonstrated for coherence in three out of the four epileptic and in both non-epileptic rats, while for STLmax in all four epileptic and in both non-epileptic rats. This study provided us with the opportunity to quantitatively investigate several aspects of epileptogenesis following traumatic brain injury. Our results strongly support a network pathology that worsens with time. It is conceivable that the observed changes in spatiotemporal dynamics after an initial brain insult, and long before the development of epilepsy, could constitute a basis for predictors of epileptogenesis in TBI patients.
ContributorsTobin, Edward (Author) / Iasemidis, Leonidas (Thesis advisor) / Tsakalis, Konstantinos (Committee member) / Muthuswamy, Jitendran (Committee member) / Arizona State University (Publisher)
Created2012
Description
This thesis examines themodeling, analysis, and control system design issues for scramjet powered hypersonic vehicles. A nonlinear three degrees of freedom longitudinal model which includes aero-propulsion-elasticity effects was used for all analyses. This model is based upon classical compressible flow and Euler-Bernouli structural concepts. Higher fidelity computational fluid dynamics and finite element methods are needed for more precise intermediate and final evaluations. The methods presented within this thesis were shown to be useful for guiding initial control relevant design. The model was used to examine the vehicle's static and dynamic characteristics over the vehicle's trimmable region. The vehicle has significant longitudinal coupling between the fuel equivalency ratio (FER) and the flight path angle (FPA). For control system design, a two-input two-output plant (FER - elevator to speed-FPA) with 11 states (including 3 flexible modes) was used. Velocity, FPA, and pitch were assumed to be available for feedback. Aerodynamic heat modeling and design for the assumed TPS was incorporated to original Bolender's model to study the change in static and dynamic properties. De-centralized control stability, feasibility and limitations issues were dealt with the change in TPS elasticity, mass and physical dimension. The impact of elasticity due to TPS mass, TPS physical dimension as well as prolonged heating was also analyzed to understand performance limitations of de-centralized control designed for nominal model.
ContributorsKhatri, Jaidev (Author) / Rodriguez, Armando Antonio (Thesis advisor) / Tsakalis, Konstantinos (Committee member) / Wells, Valana (Committee member) / Arizona State University (Publisher)
Created2011
Description
The field of education has been immensely benefited by major breakthroughs in technology. The arrival of computers and the internet made student-teacher interaction from different parts of the world viable, increasing the reach of the educator to hitherto remote corners of the world. The arrival of mobile phones in the recent past has the potential to provide the next paradigm shift in the way education is conducted. It combines the universal reach and powerful visualization capabilities of the computer with intimacy and portability. Engineering education is a field which can exploit the benefits of mobile devices to enhance learning and spread essential technical know-how to different parts of the world. In this thesis, I present AJDSP, an Android application evolved from JDSP, providing an intuitive and a easy to use environment for signal processing education. AJDSP is a graphical programming laboratory for digital signal processing developed for the Android platform. It is designed to provide utility; both as a supplement to traditional classroom learning and as a tool for self-learning. The architecture of AJDSP is based on the Model-View-Controller paradigm optimized for the Android platform. The extensive set of function modules cover a wide range of basic signal processing areas such as convolution, fast Fourier transform, z transform and filter design. The simple and intuitive user interface inspired from iJDSP is designed to facilitate ease of navigation and to provide the user with an intimate learning environment. Rich visualizations necessary to understand mathematically intensive signal processing algorithms have been incorporated into the software. Interactive demonstrations boosting student understanding of concepts like convolution and the relation between different signal domains have also been developed. A set of detailed assessments to evaluate the application has been conducted for graduate and senior-level undergraduate students.
ContributorsRanganath, Suhas (Author) / Spanias, Andreas (Thesis advisor) / Tepedelenlioğlu, Cihan (Committee member) / Tsakalis, Konstantinos (Committee member) / Arizona State University (Publisher)
Created2013
Description
Localization tasks using two-way ranging (TWR) are making headway in modern daynavigation applications as an alternative to legacy global navigation satellite systems
(GNSS) such as GPS. There is not currently literature that provides a closed-form
expression for estimation performance bounds on position and attitude when a TWR
system is employed. A Cramer-Rao Lower Bounds (CRLB) is derived for position
and orientation estimation using both 2-D and 3-D geometries. A literature review is
performed to give background and detail on the tools needed for a thorough analysis of
this problem. Popular Least Squares techniques and solutions to Wahba’s problem are
compared to the derived bounds as proof of correctness using Monte Carlo simulations.
A brief exploration on estimation performance using an Extended Kalman Filter for
non-stationary users is also looked at as an introduction to future extensions to this
work. The literature Applications like the CHP2 system are discussed as well to show
how secure, inexpensive and robust implementation of TWR is highly feasible.
i
ContributorsWelker, Samuel (Author) / Bliss, Daniel (Thesis advisor) / Herschfelt, Andrew (Committee member) / Tsakalis, Konstantinos (Committee member) / Arizona State University (Publisher)
Created2022