Browsing by Subject "Redundancy"

Now showing 1 - 8 of 8

Determining how noise and task redundancy influence motor control of planar reaching
(2013-12) Nguyen, Hung Phuc, active 2013; Dingwell, Jonathan B.; Seepersad, Carolyn
Motor noise and redundancy are vexing issues in motor control; yet their understanding provides great insights on underlying control mechanisms that govern movement. They provide glimpses into how the nervous system organizes and regulates movement within the motor control system. Understand of motor control could spur new advances in motor control could lead to better development of rehabilitation process and technology to counteract debilitating affects of neuromuscular disorders and motor readjustment with prostheses. However, before such process and technology could be developed and adapted for clinical use, a deeper understanding of motor control is needed to unravel the neural roadmap that regulates and generates movement. New theory of motor control could precipitate the development of more robust control mechanisms for robotic-human interaction. This work aims at expanding a more rigorous analytical and mathematical framework to understand how these control mechanisms reconcile redundancy and stochastic noise in human motor control.
Deterministic and probabilistic analyses of offshore pile systems
(2016-08) Chen, Jinbo, Ph. D.; Gilbert, Robert B. (Robert Bruce), 1965-; Manuel, Lance; Kallivokas , Loukas F.; Cox , Brady; Murff , James Don
The offshore pile system capacity and the pile capacity model biases are important aspects in the assessment of existing offshore platforms and in the performance reliability that is achieved using the state of practice. The objectives of this research are to improve understanding of the pile system behavior, to calibrate the pile system capacity model bias factors, and to evaluate the reliabilities of offshore pile systems. A simplified single pile failure surface in terms of three dimensional pile head loads is proposed based on the analytical lower and upper solutions, and is verified through finite element analyses. Numerical lower and upper bound models are then proposed for the ultimate capacity of a pile system, and are shown to be efficient and be effective in considering global torsion and out-of-plane failures. The evidence from the survival of offshore platforms indicates that (1) well conductors should be included in assessing the pile system ultimate capacity; (2) static p-y curves should be used which increases the pile system lateral capacity by 10 to 20%; (3) the mean value of the steel yield strength should be used; (4) jacket leg stubs should be included; and (5) site-specific geotechnical information is important. The model bias factors in the API load and resistance design recipe are calibrated through Bayes’ Theorem based on the predicted and observed performance of eighteen offshore platforms in recent Gulf of Mexico hurricanes. The API load and resistance design recipe is calibrated to be close to unbiased for predicting the jacket system performance; be slightly conservative for predicting a foundation overturning failure in clay; and be conservative for predicting a lateral failure in clay and a foundation overturning failure in sand. The reliability of a pile system is shown to be insensitive to water depths and locations in the Gulf of Mexico, but depends on the pile layout, number of piles, loading direction, and expected failure mode. The pile system redundancy (a measure of capacity beyond failure of the first element) and robustness (a measure of capacity when the system is damaged) depend on the failure mode, pile geometry and layout, and loading directions. In general, the 8-leg pile system is more redundant and more robust than the 3-leg and 4-leg pile systems. The complexity (a measure of the how well the most critically-loaded element represents all elements) depends on the pile layout, the expected failure mode of a single pile and the pile capacity uncertainty. The complexity is generally small, indicating that the failure probability of the most critically-loaded pile is representative of the failure probabilities for all piles.
Experimental testing of steel box straddle caps for evaluation of internal redundancy
(2022-05-05) Williams, Emma C., 1998-; Helwig, Todd Aaron, 1965-; Engelhardt, Michael D.
The work outlined in this thesis is a part of a larger study on the development of details to provide internal redundancy to steel box straddle caps which are currently classified as fracture critical members. This study is funded by the Texas Department of Transportation (TxDOT 0-7012). The fundamental goal of this investigation is to develop and evaluate box section details which will no longer carry the stringent inspection requirements of fracture critical members. Additionally, researchers will produce design examples for the implementation of these details. This work includes laboratory testing and parametric FEA studies. The experimental testing involves evaluating the performance of the selected details during the propagation of a brittle fracture and the capacity of the system after the fracture has occurred. The data from the experimental testing will be used to validate parametric FEA models. These models will be used to evaluate the behavior of various arrangements and applications of the developed details. This thesis focuses on the experimental testing of the small-scale pilot specimens and the large-scale baselines specimen, which possessed no detailing for redundancy. It includes the documentation of the fatigue crack growth and brittle fracture propagation procedures implemented on this study.
RDDR : n-versioning of microservices
(2020-05-08) Wood, Riley Jacob; Tiwari, Mohit
N-versioning is a well-studied method to increase the reliability of software. In this paper, we study n-versioning as applied to microservice-based applications. We construct a generic proxy called RDDR that orchestrates and monitors N variants of a microservice in order to detect bugs that make them behave differently. We showcase RDDR's ability to close five exemplary information leaks, where diversity is derived from: different software versions, different implementations of the same logical service, and variation provided by the OS like ASLR. These case studies feature information leakage through both frontend and backend interfaces of various web applications. To show that RDDR can close vulnerabilities while handling large volumes of benign tra ffic, we also apply RDDR to components of GitLab, a complex cloud application. Finally, we quantify the performance overhead associated with deploying RDDR. Our findings indicate that RDDR can patch information leaks while incurring approximately 3x CPU and memory overhead for a deployment with 3 redundant instances as expected, with modest impact to throughput and latency.
System reliability for bridge bent foundations on driven piles
(2013-05) Hall, Taylor Quinn; Gilbert, Robert B. (Robert Bruce), 1965-
Modern day bridge foundations are often designed for the maximum loading condition on a single member and each following foundation member often follows the same design. A 4-pile and a 2-pile foundation system were analyzed with an upperbound plasticity model in an attempt to approximately represent the system capacity. A sensitivity analysis was performed to measure the change in the ultimate system capacity by altering the parameters used to define individual pile capacities. The load and the resistance are considered to be normally distributed and are each characterized by a mean and a standard deviation. A first order reliability method was used to measure and account for any system redundancy in the foundation system by predicting a probability of failure based on the load and resistance of the system. Based on the most heavily loaded pile and the analysis performed on the 2-pile and 4-pile foundation systems, the probabilities of failure are lower for the system than for the single pile. This would bring reason to state there is some measurable redundancy in the basic systems analyzed where these systems would once be considered to be non-redundant systems.
Techniques to minimize circuitry and improve efficiency for defect tolerance
(2013-05) Rab, Muhammad Tauseef; Touba, Nur A.
As technology continues to scale to smaller geometries and newer dimensions (3-D), with increasingly complex manufacturing processes, the ability to reliably manufacture 100% defect-free circuitry becomes a significant challenge. While implementing additional circuitry to improve yield is economically justifiable, this thesis addresses the cost of defect tolerance by providing lower cost solutions or alternatively more defect tolerance for the same cost in state-of-the-art ICs, including three-dimensional ICs (3-D ICs). Conventional defect tolerance techniques involve incorporating redundancy into the design. This thesis introduces novel designs to maximize the utility of spare elements with minimal circuitry overhead, thereby improving the yield. One idea proposed is Selective Row Partitioning (SRP), a technique which allows a single spare column to be used to repair multiple defective cells in multiple columns. This is done by selectively decoding the row address bits when generating the select signals for the column multiplexers. This logically segments the spare column allowing it to replace different columns in different partitions of the row address space. All the chips are identical, but fuses are used to customize the row decoding circuitry on a chip-by-chip basis. An implementation procedure and results are presented which show improvement in overall yield at a minimal overhead cost. Moreover, new yield-enhancing design techniques for 3-D ICs are introduced. When assembling a 3-D IC, there are several degrees of freedom including which die are stacked together, in what order, and with what rotational symmetry. This thesis describes strategies for exploiting these degrees of freedom to reduce the cost and complexity of implementing defect tolerance. One strategy is to enable asymmetric repair capability within a 3-D memory stack by exploiting the degree of freedom that the order of the die in the stack can be selected. This technique optimizes the number of fuses, and in some cases, the number of spares as well, required to implement defect tolerance. Another innovative technique is to exploit rotational symmetry of the dies to do implicit reconfiguration to implement defect tolerance. Results show that leakage power and performance overhead for defect tolerance can be significantly reduced by this technique.
Trial-to-trial dynamics and learning in generalized, redundant reaching tasks
(2010-08) Smallwood, Rachel Fay; Dingwell, Jonathan B.; Abraham, Lawrence D.
Trial-to-trial variability in human movement is often overlooked and averaged out, but useful information can be gleaned on the brain’s control of variability. A task can be defined by a function specifying a solution manifold along which all task variable combinations will lead to goal success – the Goal-Equivalent Manifold (GEM). We selected a reaching task with variables reach Distance (D) and reach Time (T). Two GEMs were selected: a constant D/T and constant D×T. Subjects had no knowledge of the goal prior to the experiments and were instructed only to minimize error. Subjects learned the generalized tasks by reducing errors and consolidated learning from one day to the next, generalized learning from the D×T to the D/T GEM, and had interference of learning from the D/T to the D×T GEM. Variability was structured along each GEM significantly more than perpendicular to it. Deviations resulting in errors were corrected significantly more quickly than any other deviation. Our results indicate that subjects can learn generalized reaching tasks, and the brain exploits redundancy in those tasks.
Two-dimensional models of goal-oriented trial-to-trial error correction dynamics for a redundant goal : a constructive comparison
(2016-08) Devine, Mary Rose; Dingwell, Jonathan B.
Human movements are variable, even in well-learned, controlled tasks of repeated movements. Simple models of repeated movements help us understand how the control of movements and the inherent noise in a system interact and influence the measurable variability in the outcome movements (the task). Here, we compare contemporary models for correcting repeated movements in the presence of noise, with a redundant goal (i.e. one that has many solutions) in the two dimensional task space. We show that the models share a similar structure, and explain their differences in noise processes. We compare simulations of model behavior to data from a previously published reaching task, to understand what features of the models we need in general. Ultimately, our simulations show that the correction or controller with free parameters in two independent directions is necessary to describe two-dimensional tasks in general. However, we cannot conclude in favor of one model over the other, because simulations also show that either of the different noise processes is sufficient.