Digital Commons Network^™

Using Automated Task Solution Synthesis To Generate Critical Junctures For Management Of Planned And Reactive Cooperation Between A Human-Controlled Blimp And An Autonomous Ground Robot, Christopher M. Reardon

Masters Theses

This thesis documents the use of an approach for automated task solution synthesis that algorithmically and automatically identifies periods during which a team of less-than-fully capable robots benefit from tightly-coupled, coordinated, cooperative behavior.

I test two hypotheses: 1) That a team’s performance can be increased by cooperating during certain specific periods of a mission and 2) That these periods can be identified automatically and algorithmically. I also demonstrate how identification of cooperative periods can be performed both off-line prior to the application and reactively during mission execution.

I validate these premises in a real-world experiment using a human-piloted Unmanned Aerial …

L-Encoder: Video Transcoding In The Logistical Network, Harold Thomas Gonzales

Masters Theses

Transcoding, the transformation of digital information from one encoding format to another, is a prominent operation in the realm of digital video and audio. This process of changing the encoding formats of multimedia files and streams is now a common task for many users due in part to the prevalence of portable media players and consumer electronic devices for digital media, which place constraints on the file and stream formats they can play. The transcoding process is often both data and compute intensive due to both the large data requirements of modern media formats from sources such as HDTV and …

Improving Query Efficiency Of Multiresolution Data Aggregation For Analytical Exploration Of Large Relational Dataset, Yu-Ting Huang

Masters Theses

Tree structures are developed to organize the data more efficiently by using fewer resources for searching records. This thesis investigates the tree's actual structure on the hard disk to see the influence of different kinds of structures on the tree building time and query execution time. In order to improve efficiency, one multi-file structure and two reorganization structures, BFS and DFS, are examined to achieve this goal. The results indicate that two reorganization structures on KDB-tree structure can significantly improve the query execution time in large data sets and multi-file structure can slightly improve the TPIE lover level building time.

Image Analysis Techniques For Vertebra Anomaly Detection In X-Ray Images, Mohammed Das

Masters Theses

"In this research, imaging techniques are investigated for the analysis and detection of abnormalities in cervical and lumbar vertebrae. Detecting vertebra anomalies pertaining to osteoarthritis such as claw, traction and anterior osteophytes can aide in treatment plans for the patient. New size invariant features were developed for the detection of claw, traction and anterior osteophytes in cervical spine vertebrae. Using a K-means clustering and nearest centroid classification approach, the results were generated that were capable of discriminating cervical vertebrae for presence of anomalies related to osteophytes. The techniques developed can be integrated into systems based on querying spine images to …

An Economic Incentive Based Routing Protocol Incorporating Quality Of Service For Mobile Peer-To-Peer Networks, Anil Jade

Masters Theses

"Economic incentive models are becoming increasingly popular in Mobile Peer to Peer Networks (M-P2P). These models entice node participation to combat free riding and to effectively manage constraint resources in the network. Due to the dynamic topology of the M-P2P network, the connections between the peers become unpredictable and therefore, reliable routing becomes important. Many routing protocols proposed earlier (such as DSR, AODV) are based on best effort data traffic policy, such as the shortest route selection (hop minimization). Using economic models to find a cost effective optimal route from the source to the destination, while considering Quality of Service …

Co-Evolutionary Automated Software Correction: A Proof Of Concept, Joshua Lee Wilkerson

Masters Theses

"The task of ensuring that a software artifact is correct can be a very time consuming process. To be able to say that an algorithm is correct is to say that it will produce results in accordance with its specifications for all valid input. One possible way to identify an incorrect implementation is through the use of automated testing (currently an open problem in the field of software engineering); however, actually correcting the implementation is typically a manual task for the software developer. In this thesis a system is presented which automates not only the testing but also the correction …

Cross-Layer Design Through Joint Routing And Link Allocation In Wireless Sensor Networks, Xuan Gong

Masters Theses

"Both energy and bandwidth are scarce resources in sensor networks. In the past, the energy efficient routing problem has been extensively studied in efforts to maximize sensor network lifetimes, but the link bandwidth has been optimistically assumed to be abundant. Because energy constraint affects how data should be routed, link bandwidth affects not only the routing topology, but also the allowed data rate on each link, which in turn affects the lifetime. Previous research that focus on energy efficient operations in sensor networks with the sole objective of maximizing network lifetime only consider the energy constraint ignoring the bandwidth constraint. …

A Time Series Classifier, Christopher Mark Gore

Masters Theses

"A time series is a sequence of data measured at successive time intervals. Time series analysis refers to all of the methods employed to understand such data, either with the purpose of explaining the underlying system producing the data or to try to predict future data points in the time series...An evolutionary algorithm is a non-deterministic method of searching a solution space, and modeled after biological evolutionary processes. A learning classifier system (LCS) is a form of evolutionary algorithm that operates on a population of mapping rules. We introduce the time series classifier TSC, a new type of LCS that …

Co-Optimization: A Generalization Of Coevolution, Travis Service

Masters Theses

"Many problems encountered in computer science are best stated in terms of interactions amongst individuals. For example, many problems are most naturally phrased in terms of finding a candidate solution which performs best against a set of test cases. In such situations, methods are needed to find candidate solutions which are expected to perform best over all test cases. Coevolution holds the promise of addressing such problems by employing principles from biological evolution, where populations of candidate solutions and test cases are evolved over time to produce higher quality solutions...This thesis presents a generalization of coevolution to co-optimization, where optimization …

Critical Infrastructure Protection And The Domain Name Service (Dns) System, Mark Edward Snyder

Masters Theses

"Components of the critical infrastructure of any system are natural targets for attack. Any inherent weakness of such components can potentially expose the entire system to vulnerability. The Domain Name System (DNS) is one component of the proper functioning of the Internet. Although DNS is a relatively simple, isolated component, it serves as a straightforward example for the study of distributed systems in general, and as such, we have explored properties of DNS to examine how enterprise-scale, critical infrastructure components are vulnerable to attack, what protections are afforded to defenders of such components, the inherent weaknesses of such systems, and …

Preserd - Privacy Ensured Service Discovery In Mobile Peer-To-Peer Environment, Santhosh Muthyapu

Masters Theses

"In mobile peer-to-peer networks many service discovery protocols have been proposed. Most of these protocols disregard the exposure of the participating peers' privacy details, although they consider the security issues. In these methods, the participating peers must provide their identities, during the service discovery process, to be authorized to utilize the service. However, a peer might not be willing to reveal its identity until it identifies the service providing peer. So these peers face a problem; should the requesting peer or the service providing peer reveal the identity first, and hence, this is similar to the chicken-and-egg problem. The protocol …