Voice over IP (VoIP) is a key enabling technology for the migration of circuit-switched PSTN architectures to packet-based IP networks. However, this migration is successful only if the present problems in IP networks are addressed before deploying VoIP infrastructure on a large scale. One of the important issues that the present VoIP networks face is the problem of unwanted calls commonly referred to as SPIT (spam over Internet telephony). Mostly, these SPIT calls are from unknown callers who broadcast unwanted calls. There may be unwanted calls from legitimate and known people too. In this case, the unwantedness depends on social proximity of the communicating parties. For detecting these unwanted calls, I propose a framework that analyzes incoming calls for unwanted behavior. The framework includes a VoIP spam detector (VSD) that analyzes incoming VoIP calls for spam behavior using trust and reputation techniques. The framework also includes a nuisance detector (ND) that proactively infers the nuisance (or reluctance of the end user) to receive incoming calls. This inference is based on past mutual behavior between the calling and the called party (i.e., caller and callee), the callee's presence (mood or state of mind) and tolerance in receiving voice calls from the …

With the advances in very large scale integration (VLSI) technology, hundreds of billions of transistors can be packed into a single chip. With the increased hardware budget, how to take advantage of available hardware resources becomes an important research area. Some researchers have shifted from control flow Von-Neumann architecture back to dataflow architecture again in order to explore scalable architectures leading to multi-core systems with several hundreds of processing elements. In this dissertation, I address how the performance of modern processing systems can be improved, while attempting to reduce hardware complexity and energy consumptions. My research described here tackles both central processing unit (CPU) performance and memory subsystem performance. More specifically I will describe my research related to the design of an innovative decoupled multithreaded architecture that can be used in multi-core processor implementations. I also address how memory management functions can be off-loaded from processing pipelines to further improve system performance and eliminate cache pollution caused by runtime management functions.

Current syndromic surveillance systems utilize centralized databases that are neither scalable in storage space nor in computing power. Such systems are limited in the amount of syndromic data that may be collected and analyzed for the early detection of infectious disease outbreaks. However, with the increased prevalence of international travel, public health monitoring must extend beyond the borders of municipalities or states which will require the ability to store vasts amount of data and significant computing power for analyzing the data. Intelligent mobile agents may be used to create a distributed surveillance system that will utilize the hard drives and computer processing unit (CPU) power of the hosts on the agent network where the syndromic information is located. This thesis proposes the design of a mobile agent-based syndromic surveillance system and an agent decision model for outbreak detection. Simulation results indicate that mobile agents are capable of detecting an outbreak that occurs at all hosts the agent is monitoring. Further study of agent decision models is required to account for localized epidemics and variable agent movement rates.

The majority of the world's languages are poorly represented in informational media like radio, television, newspapers, and the Internet. Translation into and out of these languages may offer a way for speakers of these languages to interact with the wider world, but current statistical machine translation models are only effective with a large corpus of parallel texts - texts in two languages that are translations of one another - which most languages lack. This thesis describes the Babylon project which attempts to alleviate this shortage by supplementing existing parallel texts with texts gathered automatically from the Web -- specifically targeting pages that contain text in a pair of languages. Results indicate that parallel texts gathered from the Web can be effectively used as a source of training data for machine translation and can significantly improve the translation quality for text in a similar domain. However, the small quantity of high-quality low-density language parallel texts on the Web remains a significant obstacle.

Recently, there is a great interest in moving object tracking in the fields of security and surveillance. Object recognition under partial occlusion is the core of any object tracking system. This thesis presents an automatic and real-time color object-recognition system which is not only robust but also occlusion tolerant. The intended use of the system is to recognize and track external vehicles entered inside a secured area like a school campus or any army base. Statistical morphological skeleton is used to represent the visible shape of the vehicle. Simple curve matching and different feature based matching techniques are used to recognize the segmented vehicle. Features of the vehicle are extracted upon entering the secured area. The vehicle is recognized from either a digital video frame or a static digital image when needed. The recognition engine will help the design of a high performance tracking system meant for remote video surveillance.

Abstract Probabilistic reasoning under uncertainty suits well to analysis of disease dynamics. The stochastic nature of disease progression is modeled by applying the principles of Bayesian learning. Bayesian learning predicts the disease progression, including prevalence and incidence, for a geographic region and demographic composition. Public health resources, prioritized by the order of risk levels of the population, will efficiently minimize the disease spread and curtail the epidemic at the earliest. A Bayesian network representing the outbreak of influenza and pneumonia in a geographic region is ported to a newer region with different demographic composition. Upon analysis for the newer region, the corresponding prevalence of influenza and pneumonia among the different demographic subgroups is inferred for the newer region. Bayesian reasoning coupled with disease timeline is used to reverse engineer an influenza outbreak for a given geographic and demographic setting. The temporal flow of the epidemic among the different sections of the population is analyzed to identify the corresponding risk levels. In comparison to spread vaccination, prioritizing the limited vaccination resources to the higher risk groups results in relatively lower influenza prevalence. HIV incidence in Texas from 1989-2002 is analyzed using demographic based epidemic curves. Dynamic Bayesian networks are integrated with …

Extensive research has been conducted by the grid community to enable large-scale collaborations in pre-configured environments. grid collaborations can vary in scale and motivation resulting in a coarse classification of grids: national grid, project grid, enterprise grid, and volunteer grid. Despite the differences in scope and scale, all the traditional grids in practice share some common assumptions. They support mutually collaborative communities, adopt a centralized control for membership, and assume a well-defined non-changing collaboration. To support grid applications that do not confirm to these assumptions, we propose the concept of ad hoc grids. In the context of this research, we propose a novel architecture for ad hoc grids that integrates a suite of component frameworks. Specifically, our architecture combines the community management framework, security framework, abstraction framework, quality of service framework, and reputation framework. The overarching objective of our integrated architecture is to support a variety of grid applications in a self-controlled fashion with the help of a self-organizing ad hoc community. We introduce mechanisms in our architecture that successfully isolates malicious elements from the community, inherently improving the quality of grid services and extracting deterministic quality assurances from the underlying infrastructure. We also emphasize on the technology-independence of our …

Abstract This dissertation investigates authentication techniques in some emerging areas. Specifically, authentication schemes have been proposed that are well-suited for embedded systems, and privacy-respecting pay Web sites. With embedded systems, a person could own several devices which are capable of communication and interaction, but these devices use embedded processors whose computational capabilities are limited as compared to desktop computers. Examples of this scenario include entertainment devices or appliances owned by a consumer, multiple control and sensor systems in an automobile or airplane, and environmental controls in a building. An efficient public key cryptosystem has been devised, which provides a complete solution to an embedded system, including protocols for authentication, authenticated key exchange, encryption, and revocation. The new construction is especially suitable for the devices with constrained computing capabilities and resources. Compared with other available authentication schemes, such as X.509, identity-based encryption, etc, the new construction provides unique features such as simplicity, efficiency, forward secrecy, and an efficient re-keying mechanism. In the application scenario for a pay Web site, users may be sensitive about their privacy, and do not wish their behaviors to be tracked by Web sites. Thus, an anonymous authentication scheme is desirable in this case. That is, a …

Language can sometimes be an impediment in communication. Whether we are talking about people who speak different languages, students who are learning a new language, or people with language disorders, the understanding of linguistic representations in a given language requires a certain amount of knowledge that not everybody has. In this thesis, we propose "translation through pictures" as a means for conveying simple pieces of information across language barriers, and describe a system that can automatically generate pictorial representations for simple sentences. Comparative experiments conducted on visual and linguistic representations of information show that a considerable amount of understanding can be achieved through pictorial descriptions, with results within a comparable range of those obtained with current machine translation techniques. Moreover, a user study conducted around the pictorial translation system reveals that users found the system to generally produce correct word/image associations, and rate the system as interactive and intelligent.

Design of systems in the low-end nanometer domain has introduced new dimensions in power consumption and dissipation in CMOS devices. With continued and aggressive scaling, using low thickness SiO2 for the transistor gates, gate leakage due to gate oxide direct tunneling current has emerged as the major component of leakage in the CMOS circuits. Therefore, providing a solution to the issue of gate oxide leakage has become one of the key concerns in achieving low power and high performance CMOS VLSI circuits. In this thesis, a new approach is proposed involving dual dielectric of dual thicknesses (DKDT) for the reducing both ON and OFF state gate leakage. It is claimed that the simultaneous utilization of SiON and SiO2 each with multiple thicknesses is a better approach for gate leakage reduction than the conventional usage of a single gate dielectric (SiO2), possibly with multiple thicknesses. An algorithm is developed for DKDT assignment that minimizes the overall leakage for a circuit without compromising with the performance. Extensive experiments were carried out on ISCAS'85 benchmarks using 45nm technology which showed that the proposed approach can reduce the leakage, as much as 98% (in an average 89.5%), without degrading the performance.

Partial order planning is an important approach that solves planning problems without completely specifying the orderings between the actions in the plan. This property provides greater flexibility in executing plans; hence making the partial order planners a preferred choice over other planning methodologies. However, in order to find partially ordered plans, partial order planners perform a search in plan space rather than in space of world states and an uninformed search in plan space leads to poor efficiency. In this thesis, I discuss applying a reinforcement learning method, called First-visit Monte Carlo method, to partial order planning in order to design agents which do not need any training data or heuristics but are still able to make informed decisions in plan space based on experience. Communicating effectively with the agent is crucial in reinforcement learning. I address how this task was accomplished in plan space and the results from an evaluation of a blocks world test bed.

This thesis presents a secure digital camera (SDC) that inserts biometric data into images found in forms of identification such as the newly proposed electronic passport. However, putting biometric data in passports makes the data vulnerable for theft, causing privacy related issues. An effective solution to combating unauthorized access such as skimming (obtaining data from the passport's owner who did not willingly submit the data) or eavesdropping (intercepting information as it moves from the chip to the reader) could be judicious use of watermarking and encryption at the source end of the biometric process in hardware like digital camera or scanners etc. To address such issues, a novel approach and its architecture in the framework of a digital camera, conceptualized as an SDC is presented. The SDC inserts biometric data into passport image with the aid of watermarking and encryption processes. The VLSI (very large scale integration) architecture of the functional units of the SDC such as watermarking and encryption unit is presented. The result of the hardware implementation of Rijndael advanced encryption standard (AES) and a discrete cosine transform (DCT) based visible and invisible watermarking algorithm is presented. The prototype chip can carry out simultaneous encryption and watermarking, which …

A local call admission control (CAC) algorithm for third generation wireless networks was designed and implemented, which allows for the simulation of network throughput for different spreading factors and various mobility scenarios. A global CAC algorithm is also implemented and used as a benchmark since it is inherently optimized; it yields the best possible performance but has an intensive computational complexity. Optimized local CAC algorithm achieves similar performance as global CAC algorithm at a fraction of the computational cost. Design of a dynamic channel assignment algorithm for IEEE 802.11 wireless systems is also presented. Channels are assigned dynamically depending on the minimal interference generated by the neighboring access points on a reference access point. Analysis of dynamic channel assignment algorithm shows an improvement by a factor of 4 over the default settings of having all access points use the same channel, resulting significantly higher network throughput.

Susceptibles-infectives-removals (SIR) and its derivatives are the classic mathematical models for the study of infectious diseases in epidemiology. In order to model and simulate epidemics of an infectious disease, a global stochastic field simulation paradigm (GSFS) is proposed, which incorporates geographic and demographic based interactions. The interaction measure between regions is a function of population density and geographical distance, and has been extended to include demographic and migratory constraints. The progression of diseases using GSFS is analyzed, and similar behavior to the SIR model is exhibited by GSFS, using the geographic information systems (GIS) gravity model for interactions. The limitations of the SIR and similar models of homogeneous population with uniform mixing are addressed by the GSFS model. The GSFS model is oriented to heterogeneous population, and can incorporate interactions based on geography, demography, environment and migration patterns. The progression of diseases can be modeled at higher levels of fidelity using the GSFS model, and facilitates optimal deployment of public health resources for prevention, control and surveillance of infectious diseases.

Hard real-time systems in robotics, space and military missions, and control devices are specified with stringent and critical time constraints. On the other hand, soft real-time applications arising from multimedia, telecommunications, Internet web services, and games are specified with more lenient constraints. Real-time systems can also be distinguished in terms of their implementation into preemptive and non-preemptive systems. In preemptive systems, tasks are often preempted by higher priority tasks. Non-preemptive systems are gaining interest for implementing soft-real applications on multithreaded platforms. In this dissertation, I propose a new algorithm that uses a two-level scheduling strategy for scheduling non-preemptive soft real-time tasks. Our goal is to improve the success ratios of the well-known earliest deadline first (EDF) approach when the load on the system is very high and to improve the overall performance in both underloaded and overloaded conditions. Our approach, known as group-EDF (gEDF), is based on dynamic grouping of tasks with deadlines that are very close to each other, and using a shortest job first (SJF) technique to schedule tasks within the group. I believe that grouping tasks dynamically with similar deadlines and utilizing secondary criteria, such as minimizing the total execution time can lead to new and more …

Natural language interfaces to databases (NLIDB) are systems that aim to bridge the gap between the languages used by humans and computers, and automatically translate natural language sentences to database queries. This thesis proposes a novel approach to NLIDB, using graph-based models. The system starts by collecting as much information as possible from existing databases and sentences, and transforms this information into a knowledge base for the system. Given a new question, the system will use this knowledge to analyze and translate the sentence into its corresponding database query statement. The graph-based NLIDB system uses English as the natural language, a relational database model, and SQL as the formal query language. In experiments performed with natural language questions ran against a large database containing information about U.S. geography, the system showed good performance compared to the state-of-the-art in the field.

Due to the long duration required to perform manual knowledge entry by human knowledge engineers it is desirable to find methods to automatically acquire knowledge about the world by accessing online information. In this work I examine using the Cyc ontology to guide the creation of Naïve Bayes classifiers to provide knowledge about items described in Wikipedia articles. Given an initial set of Wikipedia articles the system uses the ontology to create positive and negative training sets for the classifiers in each category. The order in which classifiers are generated and used to test articles is also guided by the ontology. The research conducted shows that a system can be created that utilizes statistical text classification methods to extract information from an ad-hoc generated information source like Wikipedia for use in a formal semantic ontology like Cyc. Benefits and limitations of the system are discussed along with future work.

In this thesis two collaborative localization techniques are studied: multidimensional scaling (MDS) and maximum likelihood estimator (MLE). A synthesis of a new location estimation method through a serial integration of these two techniques, such that an estimate is first obtained using MDS and then MLE is employed to fine-tune the MDS solution, was the subject of this research using various simulation and experimental studies. In the simulations, important issues including the effects of sensor node density, reference node density and different deployment strategies of reference nodes were addressed. In the experimental study, the path loss model of indoor environments is developed by determining the environment-specific parameters from the experimental measurement data. Then, the empirical path loss model is employed in the analysis and simulation study of the performance of collaborative localization techniques.

Modern high performance computing is dependent on parallel processing systems. Most current benchmarks reveal only the high level computational throughput metrics, which may be sufficient for single processor systems, but can lead to a misrepresentation of true system capability for parallel systems. A new benchmark is therefore proposed. CLUE (Cluster Evaluator) uses a cellular automata algorithm to evaluate the scalability of parallel processing machines. The benchmark also uses algorithmic variations to evaluate individual system components' impact on the overall serial fraction and efficiency. CLUE is not a replacement for other performance-centric benchmarks, but rather shows the scalability of a system and provides metrics to reveal where one can improve overall performance. CLUE is a new benchmark which demonstrates a better comparison among different parallel systems than existing benchmarks and can diagnose where a particular parallel system can be optimized.

The major goal of information integration is to provide efficient and easy-to-use access to multiple heterogeneous data sources with a single query. At the same time, one of the current trends is to use standard technologies for implementing solutions to complex software problems. In this dissertation, I used XML and XQuery as the standard technologies and have developed an extended projection algorithm to provide a solution to the information integration problem. In order to demonstrate my solution, I implemented a prototype mediation system called Omphalos based on XML related technologies. The dissertation describes the architecture of the system, its metadata, and the process it uses to answer queries. The system uses XQuery expressions (termed metaqueries) to capture complex mappings between global schemas and data source schemas. The system then applies these metaqueries in order to rewrite a user query on a virtual global database (representing the integrated view of the heterogeneous data sources) to a query (termed an outsourced query) on the real data sources. An extended XML document projection algorithm was developed to increase the efficiency of selecting the relevant subset of data from an individual data source to answer the user query. The system applies the projection algorithm …

I present a time synchronization algorithm for wireless sensor networks that aims to conserve sensor battery power. The proposed method creates a hierarchical tree by flooding the sensor network from a designated source point. It then uses a hybrid algorithm derived from the timing-sync protocol for sensor networks (TSPN) and the reference broadcast synchronization method (RBS) to periodically synchronize sensor clocks by minimizing energy consumption. In multi-hop ad-hoc networks, a depleted sensor will drop information from all other sensors that route data through it, decreasing the physical area being monitored by the network. The proposed method uses several techniques and thresholds to maintain network connectivity. A new root sensor is chosen when the current one's battery power decreases to a designated value. I implement this new synchronization technique using Matlab and show that it can provide significant power savings over both TPSN and RBS.

Wireless sensor networks are battery-powered ad-hoc networks in which sensor nodes that are scattered over a region connect to each other and form multi-hop networks. These nodes are equipped with sensors such as temperature sensors, pressure sensors, and light sensors and can be queried to get the corresponding values for analysis. However, since they are battery operated, care has to be taken so that these nodes use energy efficiently. One of the areas in sensor networks where an energy analysis can be done is routing. This work explores grid-based coordinated routing in wireless sensor networks and compares the energy available in the network over time for different grid sizes.

One of the problems sensor networks face is adversaries corrupting nodes along the path to the base station. One way to reduce the effect of these attacks is multipath routing. This introduces some intrusion-tolerance in the network by way of redundancy but at the cost of a higher power consumption by the sensor nodes. Erasure coding can be applied to this scenario in which the base station can receive a subset of the total data sent and reconstruct the entire message packet at its end. This thesis uses two commonly used encodings and compares their performance with respect to power consumed for unencoded data in multipath routing. It is found that using encoding with multipath routing reduces the power consumption and at the same time enables the user to send reasonably large data sizes. The experiments in this thesis were performed on the Tiny OS platform with the simulations done in TOSSIM and the power measurements were taken in PowerTOSSIM. They were performed on the simple radio model and the lossy radio model provided by Tiny OS. The lossy radio model was simulated with distances of 10 feet, 15 feet and 20 feet between nodes. It was found that by …

Placement is one of the most important steps in physical design for VLSI circuits. For field programmable gate arrays (FPGAs), the placement step determines the location of each logic block. I present novel timing and congestion driven placement algorithms for FPGAs with minimal runtime overhead. By predicting the post-routing timing-critical edges and estimating congestion accurately, this algorithm is able to simultaneously reduce the critical path delay and the minimum number of routing tracks. The core of the algorithm consists of a criticality-history record of connection edges and a congestion map. This approach is applied to the 20 largest Microelectronics Center of North Carolina (MCNC) benchmark circuits. Experimental results show that compared with the state-of-the-art FPGA place and route package, the Versatile Place and Route (VPR) suite, this algorithm yields an average of 8.1% reduction (maximum 30.5%) in the critical path delay and 5% reduction in channel width. Meanwhile, the average runtime of the algorithm is only 2.3X as of VPR.