Repository logo

Browsing by Author "Jayasumana, Anura P., advisor"

Now showing 1 - 16 of 16
  • Thumbnail Image
    ItemOpen Access
    A graph-based, systems approach for detecting violent extremist radicalization trajectories and other latent behaviors
    (Colorado State University. Libraries, 2017) Hung, Benjamin W. K., author; Jayasumana, Anura P., advisor; Chong, Edwin K. P., committee member; Ray, Indrajit, committee member; Sega, Ronald M., committee member
    The number and lethality of violent extremist plots motivated by the Salafi-jihadist ideology have been growing for nearly the last decade in both the U.S and Western Europe. While detecting the radicalization of violent extremists is a key component in preventing future terrorist attacks, it remains a significant challenge to law enforcement due to the issues of both scale and dynamics. Recent terrorist attack successes highlight the real possibility of missed signals from, or continued radicalization by, individuals whom the authorities had formerly investigated and even interviewed. Additionally, beyond considering just the behavioral dynamics of a person of interest is the need for investigators to consider the behaviors and activities of social ties vis-à-vis the person of interest. We undertake a fundamentally systems approach in addressing these challenges by investigating the need and feasibility of a radicalization detection system, a risk assessment assistance technology for law enforcement and intelligence agencies. The proposed system first mines public data and government databases for individuals who exhibit risk indicators for extremist violence, and then enables law enforcement to monitor those individuals at the scope and scale that is lawful, and account for the dynamic indicative behaviors of the individuals and their associates rigorously and automatically. In this thesis, we first identify the operational deficiencies of current law enforcement and intelligence agency efforts, investigate the environmental conditions and stakeholders most salient to the development and operation of the proposed system, and address both programmatic and technical risks with several initial mitigating strategies. We codify this large effort into a radicalization detection system framework. The main thrust of this effort is the investigation of the technological opportunities for the identification of individuals matching a radicalization pattern of behaviors in the proposed radicalization detection system. We frame our technical approach as a unique dynamic graph pattern matching problem, and develop a technology called INSiGHT (Investigative Search for Graph Trajectories) to help identify individuals or small groups with conforming subgraphs to a radicalization query pattern, and follow the match trajectories over time. INSiGHT is aimed at assisting law enforcement and intelligence agencies in monitoring and screening for those individuals whose behaviors indicate a significant risk for violence, and allow for the better prioritization of limited investigative resources. We demonstrated the performance of INSiGHT on a variety of datasets, to include small synthetic radicalization-specific data sets, a real behavioral dataset of time-stamped radicalization indicators of recent U.S. violent extremists, and a large, real-world BlogCatalog dataset serving as a proxy for the type of intelligence or law enforcement data networks that could be utilized to track the radicalization of violent extremists. We also extended INSiGHT by developing a non-combinatorial neighbor matching technique to enable analysts to maintain visibility of potential collective threats and conspiracies and account for the role close social ties have in an individual's radicalization. This enhancement was validated on small, synthetic radicalization-specific datasets as well as the large BlogCatalog dataset with real social network connections and tagging behaviors for over 80K accounts. The results showed that our algorithm returned whole and partial subgraph matches that enabled analysts to gain and maintain visibility on neighbors' activities. Overall, INSiGHT led to consistent, informed, and reliable assessments about those who pose a significant risk for some latent behavior in a variety of settings. Based upon these results, we maintain that INSiGHT is a feasible and useful supporting technology with the potential to optimize law enforcement investigative efforts and ultimately enable the prevention of individuals from carrying out extremist violence. Although the prime motivation of this research is the detection of violent extremist radicalization, we found that INSiGHT is applicable in detecting latent behaviors in other domains such as on-line student assessment and consumer analytics. This utility was demonstrated through experiments with real data. For on-line student assessment, we tested INSiGHT on a MOOC dataset of students and time-stamped on-line course activities to predict those students who persisted in the course. For consumer analytics, we tested the performance on a real, large proprietary consumer activities dataset from a home improvement retailer. Lastly, motivated by the desire to validate INSiGHT as a screening technology when ground truth is known, we developed a synthetic data generator of large population, time-stamped, individual-level consumer activities data consistent with an a priori project set designation (latent behavior). This contribution also sets the stage for future work in developing an analogous synthetic data generator for radicalization indicators to serve as a testbed for INSiGHT and other data mining algorithms.
  • Thumbnail Image
    ItemOpen Access
    An outlier detection approach for PCB testing based on Principal Component Analysis
    (Colorado State University. Libraries, 2011) He, Xin, author; Jayasumana, Anura P., advisor; Malaiya, Yashwant K., committee member; Reising, Steven C., committee member
    Capacitive Lead Frame Testing, a widely used approach for printed circuit board testing, is very effective for open solder detection. The approach, however, is affected by mechanical variations during testing and by tolerances of electrical parameters of components, making it difficult to use threshold based techniques for defect detection. A novel approach is presented in this thesis for identifying boardruns that are likely to be outliers. Based on Principal Components Analysis (PCA), this approach treats the set of capacitance measurements of individual connectors or sockets in a holistic manner to overcome the measurement and component parameter variations inherent in test data. Effectiveness of the method is evaluated using measurements on different types of boards. Based on multiple analyses of different measurement datasets, the most suitable statistics for outlier detection and relative parameter values are also identified. Enhancements to the PCA-based technique using the concept of test-pin windows are presented to increase the resolution of the analysis. When applied to one test window at a time, PCA is able to detect the physical position of potential defects. Combining the basic and enhanced techniques, the effectiveness of outlier detection is improved. The PCA based approach is extended to detect and compensate for systematic variation of measurement data caused by tilt or shift of the sense plate. This scheme promises to enhance the accuracy of outlier detection when measurements are from different fixtures. Compensation approaches are introduced to correct the 'abnormal' measurements due to sense-plate variations to a 'normal' and consistent baseline. The effectiveness of this approach in the presence of the two common forms of mechanical variations is illustrated. Potential to use PCA based analysis to estimate the relative amount of tilt and shift in sense plate is demonstrated.
  • Thumbnail Image
    ItemOpen Access
    Anchor centric virtual coordinate systems in wireless sensor networks: from self-organization to network awareness
    (Colorado State University. Libraries, 2012) Dhanapala, Dulanjalie C., author; Jayasumana, Anura P., advisor; Kirby, Michael, committee member; Pezeshki, Ali, committee member; Ray, Indrakshi, committee member
    Future Wireless Sensor Networks (WSNs) will be collections of thousands to millions of sensor nodes, automated to self-organize, adapt, and collaborate to facilitate distributed monitoring and actuation. They may even be deployed over harsh geographical terrains and 3D structures. Low-cost sensor nodes that facilitate such massive scale networks have stringent resource constraints (e.g., in memory and energy) and limited capabilities (e.g., in communication range and computational power). Economic constraints exclude the use of expensive hardware such as Global Positioning Systems (GPSs) for network organization and structuring in many WSN applications. Alternatives that depend on signal strength measurements are highly sensitive to noise and fading, and thus often are not pragmatic for network organization. Robust, scalable, and efficient algorithms for network organization and reliable information exchange that overcome the above limitations without degrading the network's lifespan are vital for facilitating future large-scale WSN networks. This research develops fundamental algorithms and techniques targeting self-organization, data dissemination, and discovery of physical properties such as boundaries of large-scale WSNs without the need for costly physical position information. Our approach is based on Anchor Centric Virtual Coordinate Systems, commonly called Virtual Coordinate Systems (VCSs), in which each node is characterized by a coordinate vector of shortest path hop distances to a set of anchor nodes. We develop and evaluate algorithms and techniques for the following tasks associated with use of VCSs in WSNs: (a) novelty analysis of each anchor coordinate and compressed representation of VCSs; (b) regaining lost directionality and identifying a 'good' set of anchors; (c) generating topology preserving maps (TPMs); (d) efficient and reliable data dissemination, and boundary identification without physical information; and (f) achieving network awareness at individual nodes. After investigating properties and issues related to VCS, a Directional VCS (DVCS) is proposed based on a novel transformation that restores the lost directionality information in VCS. Extreme Node Search (ENS), a novel and efficient anchor placement scheme, starts with two randomly placed anchors and then uses this directional transformation to identify the number and placement of anchors in a completely distributed manner. Furthermore, a novelty-filtering-based approach for identifying a set of 'good' anchors that reduces the overhead and power consumption in routing is discussed. Physical layout information such as physical voids and even relative physical positions of sensor nodes with respect to X-Y directions are absent in a VCS description. Obtaining such information independent of physical information or signal strength measurements has not been possible until now. Two novel techniques to extract Topology Preserving Maps (TPMs) from VCS, based on Singular Value Decomposition (SVD) and DVCS are presented. A TPM is a distorted version of the layout of the network, but one that preserves the neighborhood information of the network. The generalized SVD-based TPM scheme for 3D networks provides TPMs even in situations where obtaining accurate physical information is not possible. The ability to restore directionality and topology-based Cartesian coordinates makes VCS competitive and, in many cases, a better alternative to geographic coordinates. This is demonstrated using two novel routing schemes in VC domain that outperform the well-known physical information-based routing schemes. The first scheme, DVC Routing (DVCR) uses the directionality recovered by DVCS. Geo-Logical Routing (GLR) is a technique that combines the advantages of geographic and logical routing to achieve higher routability at a lower cost by alternating between topology and virtual coordinate spaces to overcome local minima in the two domains. GLR uses topology domain coordinates derived solely from VCS as a better alternative for physical location information. A boundary detection scheme that is capable of identifying physical boundaries even for 3D surfaces is also proposed. "Network awareness" is a node's cognition of its neighborhood, its position in the network, and the network-wide status of the sensed phenomena. A novel technique is presented whereby a node achieves network awareness by passive listening to routine messages associated with applications in large-scale WSNs. With the knowledge of the network topology and phenomena distribution, every node is capable of making solo decisions that are more sensible and intelligent, thereby improving overall network performance, efficiency, and lifespan. In essence, this research has laid a firm foundation for use of Anchor Centric Virtual Coordinate Systems in WSN applications, without the need for physical coordinates. Topology coordinates, derived from virtual coordinates, provide a novel, economical, and in many cases, a better alternative to physical coordinates. A novel concept of network awareness at nodes is demonstrated.
  • Thumbnail Image
    ItemOpen Access
    Application-aware in-network service and data fusion frameworks for distributed adaptive sensing systems
    (Colorado State University. Libraries, 2009) Lee, Pan Ho, author; Jayasumana, Anura P., advisor
    Distributed Collaborative Adaptive Sensing (DCAS) systems are emerging for applications, such as detection and prediction of hazardous weather using a network of radars. Collaborative Adaptive Sensing of the Atmosphere (CASA) is an example of these emerging DCAS systems. CASA is based on a dense network of weather radars that operate collaboratively to detect tornadoes and other hazardous atmospheric conditions. This dissertation presents an application-aware data transport framework to meet the data distribution/processing requirements of such mission-critical sensor applications over best-effort networks. Our application-aware data transport framework consists of overlay architecture and a programming interface. The architecture enables deploying application-aware in-network services in an overlay network to allow applications to best adapt to the network conditions. The programming interface facilitates development of applications within the architectural framework. We demonstrate the efficacy of the proposed framework by considering a DCAS application. We evaluate the proposed schemes in a network emulation environment and on Planetlab, a world-wide Internet test-bed. The proposed schemes are very effective in delivering high quality data to the multiple end users under various network conditions. This dissertation also presents the design and implementation of an architectural framework for timely and accurate processing of radar data fusion algorithms. The preliminary version of the framework is used for real-time implementation of a multi-radar data fusion algorithm, the CASA network-based reflectivity retrieval algorithm. As a part of this research, a peer-to-peer (P2P) collaboration framework for multi-sensor data fusion is presented. Simulation-based results illustrate the effectiveness of the proposed P2P framework. As multi-sensor fusion applications have a stringent real-time constraint, estimation of network delay across the sensor networks is important, particularly as they affect the quality of sensor fusion applications. We develop an analytical model for multi-sensor data fusion latency for the Internet-based sensor applications. Time scale-invariant burstiness observed across the network produces excessive network latencies. The analytical model considers the network delay due to the self-similar cross-traffic and latency for data synchronization for data fusion. A comparison of the analytical model and simulation-based results show that our model provides a good estimation for the multi-sensor data fusion latency.
  • Thumbnail Image
    ItemOpen Access
    Application-aware transport services for sensor-actuator networks
    (Colorado State University. Libraries, 2007) Banka, Tarun, author; Jayasumana, Anura P., advisor; Chandrasekar, V., advisor
    Many emerging mission-critical sensor actuator network applications rely on the best-effort service provided by the Internet for data dissemination. This dissertation investigates the paradigm of application-aware networking to meet the QoS requirements of the mission-critical applications over best-effort networks that do not provide end-to-end QoS support. An architecture framework is proposed for application-aware data dissemination using overlay networks. The application-aware architecture framework enables application-aware processing at overlay nodes in the best-effort network to meet the QoS requirements of the heterogeneous end users of mission-critical sensor-actuator network applications. An application-aware congestion control protocol performs data selection and real-time scheduling of data for transmission while considering different bandwidth and data quality requirements of heterogeneous end users. A packet-marking scheme is proposed that enables application-aware selective drop and forwarding of packets at intermediate overlay nodes during network congestion to further enhance the QoS received by the end users under dynamic network conditions. Effectiveness of the transport services based on application-aware architecture framework is demonstrated by one-to-many high-bandwidth time-series radar data dissemination protocol for CASA (Collaborative Adaptive Sensing of the Atmosphere) application. Experiment results demonstrate that under similar network conditions and available bandwidth, application-aware processing at overlay nodes significantly improves the quality of the time-series radar data delivered to the end users compared to case when no such application-aware processing is performed. Moreover, it is shown that application-aware congestion control protocol is friendly to the already existing TCP cross-traffic on the network as long as bandwidth requirements of the mission-critical applications are met. Scalability analysis of application-aware congestion control protocol shows that it is able to schedule data at cumulative rates of more than 700M bps without degrading the QoS received by multiple end users.
  • Thumbnail Image
    ItemOpen Access
    Applications of inertial measurement units in monitoring rehabilitation progress of arm in stroke survivors
    (Colorado State University. Libraries, 2011) Doshi, Saket Sham, author; Jayasumana, Anura P., advisor; Malcolm, Matthew P., committee member; Pasricha, Sudeep, committee member; Malaiya, Yashwant K., committee member
    Constraint Induced Movement Therapy (CIMT) has been clinically proven to be effective in restoring functional abilities of the affected arm among stroke survivors. Current CIMT delivery method lacks a robust technique to monitor rehabilitation progress, which results in increasing costs of stroke related health care. Recent advances in the design and manufacturing of Micro Electro Mechanical System (MEMS) inertial sensors have enabled tracking human motions reliably and accurately. This thesis presents three algorithms that enable monitoring of arm movements during CIMT by means of MEMS inertial sensors. The first algorithm quantifies the affected arm usage during CIMT. This algorithm filters the arm movement data, sampled during activities of daily life (ADL), by applying a threshold to determine the duration of affected arm movements. When an activity is performed multiple times, this algorithm counts the number of repetitions performed. Current technique uses a touch/proximity sensor and a motor activity log maintained by the patient to determine CIMT duration. Affected arm motion is a direct indicator of CIMT session and hence this algorithm tracks rehabilitation progress more accurately. Actual patients' affected arm movement data analysis shows that the algorithm does activity detection with an average accuracy of >90%. Second of the three algorithms, tracking stroke rehabilitation of affected arm through histogram of distance traversed, evaluates an objective metric to assess rehabilitation progress. The objective metric can be used to compare different stroke patients based on their functional ability in affected arm. The algorithm calculates the histogram by evaluating distances traversed over a fixed duration window. The impact of this window on algorithm's performance is analyzed. The algorithm has better temporal resolution when compared with another standard objective test, box and block test (BBT). The algorithm calculates linearly weighted area under the histogram as a score to rank various patients as per their rehabilitation progress. The algorithm has better performance for patients with chronic stroke and certain degree of functional ability. Lastly, Kalman filter based motion tracking algorithm is presented that tracks linear motions in 2D, such that only one axis can experience motion at any given time. The algorithm has high (>95%) accuracy. Data representing linear human arm motion along a single axis is generated to analyze and determine optimal parameters of Kalman filter. Cross-axis sensitivity of the accelerometer limits the performance of the algorithm over longer durations. A method to identify the 1D components of 2D motion is developed and cross-axis effects are removed to improve the performance of motion tracking algorithm.
  • Thumbnail Image
    ItemOpen Access
    Automating investigative pattern detection using machine learning & graph pattern matching techniques
    (Colorado State University. Libraries, 2022) Muramudalige, Shashika R., author; Jayasumana, Anura P., advisor; Ray, Indrakshi, committee member; Kim, Ryan G., committee member; Wang, Haonan, committee member
    Identification and analysis of latent and emergent behavioral patterns are core tasks in investigative domains such as homeland security, counterterrorism, and crime prevention. Development of behavioral trajectory models associated with radicalization and tracking individuals and groups based on such trajectories are critical for law enforcement investigations, but these are hampered by sheer volume and nature of data that need to be mined and processed. Dynamic and complex behaviors of extremists and extremist groups, missing or incomplete information, and lack of intelligent tools further obstruct counterterrorism efforts. Our research is aimed at developing state-of-the-art computational tools while building on recent advances in machine learning, natural language processing (NLP), and graph databases. In this work, we address the challenges of investigative pattern detection by developing algorithms, tools, and techniques primarily aimed at behavioral pattern tracking and identification for domestic radicalization. The methods developed are integrated in a framework, Investigative Pattern Detection Framework for Counterterrorism (INSPECT). INSPECT includes components for extracting information using NLP techniques, information networks to store in appropriate databases while enabling investigative graph searches, and data synthesis via generative adversarial techniques to overcome limitations due to incomplete and sparse data. These components enable streamlining investigative pattern detection while accommodating various use cases and datasets. While our outcomes are beneficial for law enforcement and counterterrorism applications to counteract the threat of violent extremism, as the results presented demonstrate, the proposed framework is adaptable to diverse behavioral pattern analysis domains such as consumer analytics, cybersecurity, and behavioral health. Information on radicalization activity and participant profiles of interest to investigative tasks are mostly found in disparate text sources. We integrate NLP approaches such as named entity recognition (NER), coreference resolution, and multi-label text classification to extract structured information regarding behavioral indicators, temporal details, and other metadata. We further use multiple text pre-processing approaches to improve the accuracy of data extraction. Our training text datasets are intrinsically smaller and label-wise imbalanced, which hinders direct application of NLP techniques for better results. We use a transfer learning-based, pre-trained NLP model by integrating our specific datasets and achieve noteworthy improvement in information extraction. The extracted information from text sources represents a rich knowledge network of populations with various types of connections that needs to be stored, updated, and repeatedly inspected for emergence of patterns in the long term. Therefore, we utilize graph databases as the foremost storage option while maintaining the reliability and scalability of behavioral data processing. To query suspicious and vulnerable individuals or groups, we implement investigative graph search algorithms as custom stored procedures on top of graph databases while verifying the ability to operate at scale. We use datasets in different contexts to demonstrate the wide-range applicability and the enhanced effectiveness of observing suspicious or latent trends using our investigative graph searches. Investigative data by nature is incomplete and sparse, and the number of cases that may be used for training investigators or machine learning algorithms is small. This is an inherent concern in investigative and many other contexts where the data collection is tedious, available data is limited and also may be subjected to privacy concerns. Having large datasets is beneficial to social scientists and investigative authorities to enhance their skills, and to achieve more accuracy and reliability. A not so small training data volume is also essential for application of the latest machine learning techniques for improved classification and detection. In this work, we propose a generative adversarial network (GAN) based approach with novel feature mapping techniques to synthesize additional data from a small and sparse data set while preserving the statistical characteristics. We also compare our proposed method with two likelihood approaches. i.e., multi-variate Gaussian and regular-vine copulas. We verify the robustness of the proposed technique via a simulation and real-world datasets representing diverse domains. The proposed GAN-based data generation approach is applicable to other domains as demonstrated with two applications. Initially, we extend our data generation approach by contributing to a computer security application resulting in improved phishing websites detection with synthesized datasets. We merge measured datasets with synthesized samples and re-train models to improve the performance of classification models and mitigate vulnerability against adversarial samples. The second was related to a video traffic classification application in which to the data sets are enhanced while preserving statistical similarity between the actual and synthesized datasets. For the video traffic data generation, we modified our data generation technique to capture the temporal patterns in time series data. In this application, we integrate a Wasserstein GAN (WGAN) by using different snapshots of the same video signal with feature-mapping techniques. A trace splitting algorithm is presented for training data of video traces that exhibit higher data throughput with high bursts at the beginning of the video session compared to the rest of the session. With synthesized data, we obtain 5 - 15% accuracy improvement for classification compared to only having actual traces. The INSPECT framework is validated primarily by mining detailed forensic biographies of known jihadists, which are extensively used by social/political scientists. Additionally, each component in the framework is extensively validated with a Human-In-The-Loop (HITL) process, which improves the reliability and accuracy of machine learning models, investigative graph algorithms, and other computing tools based on feedback from social scientists. The entire framework is embedded in a modular architecture where the analytical components are implemented independently and adjustable for different requirements and datasets. We verified the proposed framework's reliability, scalability, and generalizability with datasets in different domains. This research also makes a significant contribution to discrete and sparse data generation in diverse application domains with novel generative adversarial data synthesizing techniques.
  • Thumbnail Image
    ItemOpen Access
    Cooperative defense mechanisms for detection, identification and filtering of DDoS attacks
    (Colorado State University. Libraries, 2016) Mosharraf Ghahfarokhi, Negar, author; Jayasumana, Anura P., advisor; Ray, Indrakshi, advisor; Pezeshki, Ali, committee member; Malaiya, Yashwant, committee member
    To view the abstract, please see the full text of the document.
  • Thumbnail Image
    ItemOpen Access
    Decentralized and dynamic community formation in P2P networks and performance of community based caching
    (Colorado State University. Libraries, 2015) Limo, Chepchumba Soti, author; Jayasumana, Anura P., advisor; Yang, Liuqing, committee member; Papadopoulos, Christos, committee member
    Distributed Hash Tables (DHT) are commonly used in large Peer-to-Peer networks to increase the efficiently of resolving queries. Minimizing the resource discovery time in P2P networks is highly desirable to improve system-wide performance. Distributed caching is an approach used to reduce the look-up time. File sharing P2P networks have shown that there exists nodes/users who share similar interests based on semantics, geography, etc., and a group of nodes that share similar interests are said to form a community. A Community Based Caching (CBC) algorithm where nodes make caching decisions based on personal interests is investigated. One of CBC’s major contributions is that it alleviates the issue of nodes being limited to caching resources that are popular relative to the entire network. Instead, caching decisions are primarily based on a node's community affiliations and interests. Community discovery algorithms that currently exists either need a centralized source(s) to aid in community discovery or require additional messaging and complicated computations to determine whether to join a group or not. In many cases, nodes are also limited to being members of only one community at a time. A dynamic and decentralized community discovery algorithm, Dynamic Group Discovery (DGD), is proposed. DGD also allows nodes to be members of multiple communities at the same time. DGD's behavior and performance is then evaluated in conjunction with the Community Based Caching algorithm. To aid in group discovery during run time (i.e., dynamically), DGD uses special keys with embedded group identification information. Oversim, a flexible overly network simulation framework is used to evaluate the proposed DGD algorithm. Performance of DGD is compared to Chord and Static Group Allocation (SGA), in which group identification is done only once. Performance is evaluated for different network sizes, community sizes, and asymmetry among communities. Performance results are presented and analyzed when queries are resolved using cache data versus when queries are resolved using non-cache data. The analysis shows that DGD generally improves lookup performance when cache data is used to resolved queries. However, when non-cache data is used, DGD occasionally performs slightly worse than Chord and SGA. For example, in a network with 10,000 nodes, asymmetrical communities and no churn group churn, DGD outperforms Chord by approximately half a hop and 0.1 seconds in latency. When churn was introduced to the same network, DGD performance drops by approximately one hop and 0.15 seconds in latency. The results also show that approximately 90% of the queries are resolved using non-cache data and therefore, even though DGD is guaranteed to reduce lookup time when asymmetrical communities are present and cache records are to used to resolve queries, it is often not enough to significantly improve overall system performance. The results however confirm that caching resources based on personal interests really does reduced lookup performance when resolving queries using cache records.
  • Thumbnail Image
    ItemOpen Access
    Enhancing collaborative peer-to-peer systems using resource aggregation and caching: a multi-attribute resource and query aware approach
    (Colorado State University. Libraries, 2012) Bandara, H. M. N. Dilum, author; Jayasumana, Anura P., advisor; Chandrasekar, V., committee member; Massey, Daniel F., committee member; Ray, Indrajit, committee member
    To view the abstract, please see the full text of the document.
  • Thumbnail Image
    ItemOpen Access
    Extraction, characterization and modeling of network data features - a compressive sensing and robust PCA based approach
    (Colorado State University. Libraries, 2015) Bandara, Vidarshana W., author; Jayasumana, Anura P., advisor; Pezeshki, Ali, advisor; Scharf, Louis L., committee member; Ray, Indrajit, committee member; Luo, J. Rockey, committee member
    To view the abstract, please see the full text of the document.
  • Thumbnail Image
    ItemOpen Access
    Impact of resequencing buffer distribution on packet reordering
    (Colorado State University. Libraries, 2011) Mandyam Narasiodeyar, Raghunandan, author; Jayasumana, Anura P., advisor; Malaiya, Yashwant K., committee member; Pasricha, Sudeep, committee member
    Packet reordering in Internet has become an unavoidable phenomenon wherein packets get displaced during transmission resulting in out of order packets at the destination. Resequencing buffers are used at the end nodes to recover from packet reordering. This thesis presents analytical estimation methods for "Reorder Density" (RD) and "Reorder Buffer occupancy Density" (RBD) that are metrics of packet reordering, of packet sequences as they traverse through resequencing nodes with limited buffers. During the analysis, a "Lowest First Resequencing Algorithm" is defined and used in individual nodes to resequence packets back into order. The results are obtained by studying the patterns of sequences as they traverse through resequencing nodes. The estimations of RD and RBD are found to vary for sequences containing different types of packet reordering patterns such as Independent Reordering, Embedded Reordering and Overlapped Reordering. Therefore, multiple estimations in the form of theorems catering to different reordering patterns are presented. The proposed estimation models assist in the allocation of resources across intermediate network elements to mitigate the effect of packet reordering. Theorems to derive RBD from RD when only RD is available are also presented. Just like the resequencing estimation models, effective RBD for a given RD are also found to vary for different packet reordering patterns, therefore, multiple theorems catering to different patterns are presented. Such RBD estimations would be useful for allocating resources based on certain QoS criteria wherein one of the metrics is RD. Simulations driven by Internet measurement traces and random sequences are used to verify the analytical results. Since high degree of packet reordering is known to affect the quality of applications using TCP and UDP on the Internet, this study has broad applicability in the area of mobile communication and networks.
  • Thumbnail Image
    ItemOpen Access
    Top-down clustering based self-organization of collaborative wireless sensor networks
    (Colorado State University. Libraries, 2008) Bandara, H. M. N. Dilum, author; Jayasumana, Anura P., advisor; Massey, Daniel F., committee member; Chandrasekar, V., committee member
    The proposed cluster tree based routing strategy facilitates both node-to-sink and node-to-node communication. Hierarchical addresses that reflect the parent-child relationship of cluster heads is used to route data along the cluster tree. Utilization of cross-links among neighboring cluster heads and a circular path within the network approximately doubles the capacity of the network. Under ideal conditions, this approach guarantees delivery of events/queries and has a lower overhead compared to routing strategies such as rumor routing and ant routing. The cluster tree formed by our algorithm is used to identify and form Virtual Sensor networks (VSNs), an emerging concept that supports resource efficient collaborative WSNs. Our implementation of VSN is able to deliver unicast, multicast, and broadcast traffic among nodes observing similar events, efficiently. Efficacy of the VSN based approach is evaluated by simulating a subsurface chemical plume monitoring system. The algorithm is further extended to support the formation of a secure backbone that can enable secure upper layer functions and dynamic distribution of cryptographic keys, among nodes and users of collaborative sensor networks.
  • Thumbnail Image
    ItemOpen Access
    Topology inference of Smart Fabric grids - a virtual coordinate based approach
    (Colorado State University. Libraries, 2020) Pendharkar, Gayatri Arun, author; Jayasumana, Anura P., advisor; Maciejewski, Anthony A., committee member; Malaiya, Yashwant K., committee member
    Driven by increasing potency and decreasing cost/size of the electronic devices capable of sensing, actuating, processing and wirelessly communicating, the Internet of Things (IoT) is expanding into manufacturing plants, complex structures, and harsh environments with the potential to impact the way we live and work. Subnets of simple devices ranging from smart RFIDs to tiny sensors/actuators deployed in massive numbers forming complex 2-D surfaces, manifolds and complex 3-D physical spaces and fabrics will be a key constituent of this infrastructure. Smart Fabrics (SFs) are emerging with embedded IoT devices that have the ability to do things that traditional fabrics cannot, including sensing, storing, communicating, transforming data, and harvesting and conducting energy. These SFs are expected to have a wide range of applications in the near future in health monitoring, space stations, commercial building rooftops and more. With this innovative Smart Fabric technology at hand, there is a need to create algorithms for programming the smart nodes to facilitate communication, monitoring, and data routing within the fabric. Automatically detecting the location, shape, and other such physical characteristics will be essential but without resorting to localization techniques such as Global Positioning System (GPS), the size and cost of which may not be acceptable for many large-scale applications. Measuring the physical distances and obtaining geographical coordinates becomes infeasible for many IoT networks, particularly those deployed in harsh and complex environments. In SFs, the proximity between the nodes makes it impossible to deploy technology like GPS or Received Signal Strength Indicator (RSSI) for distance estimation. This thesis devises a Virtual Coordinate (VC) based method to identify the node positions and infer the shape of SFs with embedded grids of IoT devices. In various applications, we expect the nodes to communicate through randomly shaped fabrics in the presence of oddly-shaped holes. The geometry of node placement, the shape of the fabric, and dimensionality affect the identification, shape determination, and routing algorithms. The objective of this research is to infer the shape of fabric, holes, and other non-operational parts of the fabric with different grid placements. With the ability to construct the topology, efficient data routing can be achieved, damaged regions of fabric could be identified, and in general, the shape could be inferred for SFs with a wide range of sizes. Clothing and health monitoring being two essential segments of living, SFs that combines both would be a success in the textile market. SFs can be synthesized in space stations as compact sensing devices, assist in patient health monitoring, and also bring a spark to the showbiz. Identifying the position of different nodes/devices within SF grids is essential for applications and networking functions. We study and devise strategic methods for localization of SFs with rectangular grid placement of nodes using the VC approach, a viable alternative to geographical coordinates. In our system, VCs are computed using the hop distances to the anchors. For a full grid (no missing nodes), each grid node has predictable unique VCs. However, a SF grid may have holes/voids/obstacles that cause perturbations and distortion in VC pattern and may even result in non-unique VCs. Our shape inference method adaptively selects anchors from already localized nodes to compute VCs with the least amount of perturbation. We evaluate the proposed algorithm to simulate SF grids with varied sizes (i.e. number of nodes) and the number of voids. For each scenario, e.g. a SF grid with length X breadth dimensions - 19X19, 10% missing nodes, and 3 voids, we generate 60 samples of the grid with random possible placements and sizes of voids. Then, the localization algorithm is executed on these grids for all different scenarios. The final results measure the percentages of localized nodes as well as the total number of elected anchors required for the localization. We also investigate SF grids with triangular node placement and localization methods for the same. Additionally, parallelization techniques are implemented using an Message Parsing Interface (MPI) mechanism to run the simulations for rectangular and triangular grid SFs with efficient use of time and resources. To summarize, an algorithm was presented for the detection of voids in smart fabrics with embedded sensor nodes. It identifies the minimum set of node perturbations to be consistent with VCs and adaptively selects anchors to reduce uncertainty.
  • Thumbnail Image
    ItemOpen Access
    Virtual and topological coordinate based routing, mobility tracking and prediction in 2D and 3D wireless sensor networks
    (Colorado State University. Libraries, 2013) Jiang, Yi, author; Jayasumana, Anura P., advisor; Luo, J. Rockey, committee member; Ray, Indrakshi, committee member
    A Virtual Coordinate System (VCS) for Wireless Sensor Networks (WSNs) characterizes each sensor node's location using the minimum number of hops to a specific set of sensor nodes called anchors. VCS does not require geographic localization hardware such as Global Positioning System (GPS), or localization algorithms based on Received Signal Strength Indication (RSSI) measurements. Topological Coordinates (TCs) are derived from Virtual Coordinates (VCs) of networks using Singular Value Decomposition (SVD). Topology Preserving Maps (TPMs) based on TCs contain 2D or 3D network topology and directional information that are lost in VCs. This thesis extends the scope of VC and TC based techniques to 3D sensor networks and networks with mobile nodes. Specifically, we apply existing Extreme Node Search (ENS) for anchor placement for 3D WSNs. 3D Geo-Logical Routing (3D-GLR), a routing algorithm for 3D sensor networks that alternates between VC and TC domains is evaluated. VC and TC based methods have hitherto been used only in static networks. We develop methods to use VCs in mobile networks, including the generation of coordinates, for mobile sensors without having to regenerate VCs every time the topology changes. 2D and 3D Topological Coordinate based Tracking and Prediction (2D-TCTP and 3D-TCTP) are novel algorithms developed for mobility tracking and prediction in sensor networks without the need of physical distance measurements. Most existing 2D sensor networking algorithms fail or perform poorly in 3D networks. Developing VC and TC based algorithms for 3D sensor networks is crucial to benefit from the scalability, adjustability and flexibility of VCs as well as to overcome the many disadvantages associated with geographic coordinate systems. Existing ENS algorithm for 2D sensor networks plays a key role in providing a good anchor placement and we continue to use ENS algorithm for anchor selection in 3D network. Additionally, we propose a comparison algorithm for ENS algorithm named Double-ENS algorithm which uses two independent pairs of initial anchors and thereby increases the coverage of ENS anchors in 3D networks, in order to further prove if anchor selection from original ENS algorithm is already optimal. Existing Geo-Logical Routing (GLR) algorithm demonstrates very good routing performance by switching between greedy forwarding in virtual and topological domains in 2D sensor networks. Proposed 3D-GLR extends the algorithm to 3D networks by replacing 2D TCs with 3D TCs in TC distance calculation. Simulation results show that the 3D-GLR algorithm with ENS anchor placement can significantly outperform current Geographic Coordinates (GCs) based 3D Greedy Distributed Spanning Tree Routing (3D-GDSTR) algorithm in various network environments. This demonstrates the effectiveness of ENS algorithm and 3D-GLR algorithm in 3D sensor networks. Tracking and communicating with mobile sensors has so far required the use of localization or geographic information. This thesis presents a novel approach to achieve tracking and communication without geographic information, thus significantly reducing the hardware cost and energy consumption. Mobility of sensors in WSNs is considered under two scenarios: dynamic deployment and continuous movement. An efficient VC generation scheme, which uses the average of neighboring sensors' VCs, is proposed for newly deployed sensors to get coordinates without flooding based VC generation. For the second scenario, a prediction and tracking algorithm called 2D-TCTP for continuously moving sensors is developed for 2D sensor networks. Predicted location of a mobile sensor at a future time is calculated based on current sampled velocity and direction in topological domain. The set of sensors inside an ellipse-shaped detection area around the predicted future location is alerted for the arrival of mobile sensor for communication or detection purposes. Using TPMs as a 2D guide map, tracking and prediction performances can be achieved similar to those based on GCs. A simple modification for TPMs generation is proposed, which considers radial information contained in the first principle component from SVD. This modification improves the compression or folding at the edges that has been observed in TPMs, and thus the accuracy of tracking. 3D-TCTP uses a detection area in the shape of a 3D sphere. 3D-TCTP simulation results are similar to 2D-TCTP and show competence comparable to the same algorithms based on GCs although without any 3D geographic information.
  • Thumbnail Image
    ItemOpen Access
    Virtual coordinate based techniques for wireless sensor networks: a simulation tool and localization & planarization algorithms
    (Colorado State University. Libraries, 2013) Shah, Pritam, author; Jayasumana, Anura P., advisor; Pasricha, Sudeep, committee member; Malaiya, Yashwant K., committee member
    Wireless sensor Networks (WSNs) are deployments of smart sensor devices for monitoring environmental or physical phenomena. These sensors have the ability to communicate with other sensors within communication range or with a base station. Each sensor, at a minimum, comprises of sensing, processing, transmission, and power units. This thesis focuses on virtual coordinate based techniques in WSNs. Virtual Coordinates (VCs) characterize each node in a network with the minimum hop distances to a set of anchor nodes, as its coordinates. It provides a compelling alternative to some of the localization applications such as routing. Building a WSN testbed is often infeasible and costly. Running real experiments on WSNs testbeds is time consuming, difficult and sometimes not feasible given the scope and size of applications. Simulation is, therefore, the most common approach for developing and testing new protocols and techniques for sensor networks. Though many general and wireless sensor network specific simulation tools are available, no available tool currently provides an intuitive interface or a tool for virtual coordinate based simulations. A simulator called VCSIM is presented which focuses specifically on Virtual Coordinate Space (VCS) in WSNs. With this simulator, a user can easily create WSNs networks of different sizes, shapes, and distributions. Its graphical user interface (GUI) facilitates placement of anchors and generation of VCs. Localization in WSNs is important for several reasons including identification and correlation of gathered data, node addressing, evaluation of nodes' density and coverage, geographic routing, object tracking, and other geographic algorithms. But due to many constraints, such as limited battery power, processing capabilities, hardware costs, and measurement errors, localization still remains a hard problem in WSNs. In certain applications, such as security sensors for intrusion detection, agriculture, land monitoring, and fire alarm sensors in a building, the sensor nodes are always deployed in an orderly fashion, in contrast to random deployments. In this thesis, a novel transformation is presented to obtain position of nodes from VCs in rectangular, hexagonal and triangular grid topologies. It is shown that with certain specific anchor placements, a location of a node can be accurately approximated, if the length of a shortest path in given topology between a node and anchors is equal to length of a shortest path in full topology (i.e. a topology without any voids) between the same node and anchors. These positions are obtained without the need of any extra localization hardware. The results show that more than 90% nodes were able to identify their position in randomly deployed networks of 80% and 85% node density. These positions can then be used for deterministic routing which seems to have better avg. path length compared to geographic routing scheme called "Greedy Perimeter Stateless Routing (GPSR)". In many real world applications, manual deployment is not possible in exact regular rectangular, triangular or hexagonal grids. Due to placement constraint, nodes are often placed with some deviation from ideal grid positions. Because of placement tolerance and due to non-isotropic radio patterns nodes may communicate with more or less number of neighbors than needed and may form cross-links causing non-planar topologies. Extracting planar graph from network topologies is known as network planarization. Network planarization has been an important technique in numerous sensor network protocols--such as GPSR for efficient routing, topology discovery, localization and data-centric storage. Most of the present planarization algorithms are based on location information. In this thesis, a novel network planarization algorithm is presented for rectangular, hexagonal and triangular topologies which do not use location information. The results presented in this thesis show that with placement errors of up to 30%, 45%, and 30% in rectangular, triangular and hexagonal topologies respectively we can obtain good planar topologies without the need of location information. It is also shown that with obtained planar topology more nodes acquire unique VCs.