Sitharama S. Iyengar - US grants

The central purpose of this project is to enhance minority undergraduate faculty education in the areas of robotics and machine vision. This purpose is achieved by conducting an intensive training workshop on these topics for computer science faculty members selected from historically minority universities. In particular, we propose to conduct an intensive, two-week long training workshop, along with certain follow-up activities, for 15 computer science faculty members selected from minority institutions. Workshop participants are exposed to up-to-date topics in computer vision and robotics. The activities include direct interaction by Robotics Research Laboratory's faculty and extensive use of the RRL's research equipment. Also, participants are encouraged to take part in individual and group research projects based on current RRL research efforts. The final day of the workshop includes a brainstorming discussion of how the information learned in the workshop will be transferred into their curriculum. The results of the workshop will be published in a refereed journal.

Data mining is the semi-automatic discovery of patterns, associations, changes, anomalies, rules, and statistically significant structures and events in data. Thus, data mining aims to extract knowledge from sensor data, and is an emerging area with numerous commercial applications. Work in data mining ranges from theoretical work on the principles of learning and mathematical representations of data to building advanced engineering systems that perform information filtering on the web, find genes in DNA sequences, help understand trends and anomalies in economics and medicine, and detect network intrusion.

The problem gets more complicated in a distributed system where the data itself is positioned across different sites and coordination across the sites needed to mine information from the distributed data. Distributed Data Mining (DDM) arises in diverse areas such as sensor stream data generated from satellite etc.

Presently, there are several proposed ad-hoc approaches to distributed data mining (DDM) that deal with specific schemes and most of them do not address any real time analysis for solving different classes of distributed data mining problems. This project will address this very explicitly the merits of the new framework.

The Departments of Computer Science, Engineering, and Mathematics, in collaboration with the Office of Strategic Initiatives and the Center for Academic Success at Louisiana State University, is increasing the educational and career opportunities for academically talented, financially disadvantaged students to reach their fullest potential through a student/faculty partnership with a comprehensive monitoring system.

The goal of this project is to recruit financially disadvantaged students, to retain these students, and to enhance the educational experience of these students in the technology disciplines. Specific objectives are: (1) Recruit qualified students through an active K-12 educational outreach program, including math and science fairs, competitions, and in-school math and science laboratories, (2) Retain these students through the incorporation of student support and mentoring programs, (3) Enhance the educational experience of these students through workshops/seminars, industrial outreach, and community service opportunities, (4) Increase the number of undergraduates, including women and minorities in the above-mentioned disciplines, and (5) Promote intercultural participation among non-minority and minority students. In addition to the scholarships, the project consists of three activities: (1) mandatory computer and information technology training (average 2 hrs/wk), (2) mandatory academic enhancement activities (average 1 hr/wk), and (3) optional educational and industrial outreach programs (up to 3 hrs/wk). CSEMS Scholars receive additional information technology training through various workshops and training courses.

The Office of Strategic Initiatives and the Center for Academic Success, Co-PIs, and several other faculty members provide active mentoring and academic assistance for the CSEMS Scholars to maximize their chances for academic success. The faculty mentors also assist CSEMS Scholars in pursuing their professional careers though continuing graduate studies or job placement in industries relevant to their education. There are 29 students per year majoring in computer science, engineering, or mathematics disciplines who hold scholarships. Scholarships can be renewed up to graduation. CSEMS Scholars form a community that interacts regularly with each other, with faculty, and with other students on campus.

The information utility of a sensor network is dependent on the survival of critical sensor nodes whose identities are determined by the dynamics of information flow and task distribution in the network. This proposal will develop energy-efficient algorithms for dynamically identifying critical sensors and increasing sensor lifetime through optimal data transmission schedules and duty cycles. Path reliability, length and sensor energy-consumptions are three important metrics affecting network lifetime. The notion of energy weakness of information routing paths will be developed and used to derive energy-aware protocols that enhance sensor network survivability by selecting reliable information paths with low energy weakness. New techniques for measuring the vulnerability of sensor placements and task distributions will be developed and used to derive efficient tasking algorithms that maximize critical sensor survival. Public dissemination of research and software and training of graduate students will be an important component of the proposal.

Sensor networks, with their massive information acquisition potential,
represent a mechanism by which Information Technology can be used to
mitigate the vulnerabilities of society with regard to natural and
manmade catastrophic events. Two critical components of information
management in sensor networks are: (a) secure data distribution i.e.,
identifying secure in-network sensor data storage sites and determining
their specific data contents in order to maximize overall data security
and minimize routing overhead to these sites, and (b) data access via
maximally secure energy-constrained query reporting between data
nodes. These problems cannot be solved in a stand-alone manner, since
optimal solutions in one area may lead to sub-optimal solutions in
others. This work proposes the development of a unified analytical
framework using game-theoretic techniques. The proposed holistic
information management framework will enable the successful acquisition
and efficient utilization of sensornet data by identifying storage
repositories, data contents, and secure query reporting schemes that
are optimally suited to each other. Efficient algorithms for the above
problems will be derived using a new Data Security Index metric which
explicitly accounts for the actions of intelligent opponents and models
sensors as intelligent agents cooperating to achieve network objectives,
such as maximal data security and data access.

This project, developing a new kind of research instrument and tools for interactive, collaborative, networked 3D visualization, creates a new genre of physical interaction devices, called "viz tangibles." These devices allow users to load and save data, steer parameters, navigate in time, control teleconference links, and perform other key operations. The system can be easily moved between conference rooms or offices, and can be connected with mono or stereo projectors. The effort enables scientists to significantly enhance the utility, usability, and collaborative potential for scientific visualization in research, education, collaboration, and interaction with the general public. The resulting tools will be deployed, utilized, and integrated into research and education at five Louisiana campuses which are already linked together by numerous academic collaborations. The central elements of these platforms, the "viz tangibles," are small, modular interaction devices, typically embedded with different interactors, indicators, and networked microcontrollers. So far, three basic and three specialized viz tangibles, interconnected physically and functionally following a simple grammar, have been implemented. One device loads and saves online information, applications, and parameters, which are referenced with RFID-tagged cards. Another device includes a rebindable "parameter wheel," that supports the manipulation of continuous discrete parameters. A third device provides VCR-like button controls that may drive a variety of interactions. Together, these devices provide core operations for a wide variety of visualization and simulation applications, with functionality resembling the GUI "file" and "edit" menus, but optimized for simple, collaborative, semi-immersive, divided-attention use. The development includes researchers with expertise in visualization, human-computer-interaction, grid computing, software engineering, and electronic, mechanical, and product design. Collaborators include faculty in biomechanics from LSU Medical, Southern U., LaTech; chemistry; computational fluid dynamics; coastal and hurricane modeling; and numerical relativity.

Broader Impact: Application of these viz tangibles range from the laboratory to the science education center. The network links research institutions, a medical center, a historically black university, and a supercomputer center.

The Department of Computer Science, National Center for Security Research and Training, and Office of Strategic Initiatives at Louisiana State University in collaboration with Southern University and Louisiana Tech University are providing educational and pedagogical opportunities for academically talented faculty in information assurance and security.

Intellectual merit. The project, "Multi-University Research and Training in Information Assurance and Computer Security," engages faculty scholars in a series of faculty development activities that include a summer workshop, national lab/research center visits, academic year mini-grants, and a conference in information assurance and computer security. Program activities are being offered through two phases. During the first phase faculty participate in a 2-week summer workshop and national lab/research center visits that focus on state-of-the-art research training and various enhancement activities. During the second phase, faculty scholars receive academic year mini-grants that extend research and education training to the home institutions and participate in an annual conference that serves as a consolidation of experience exchange and national dissemination.

Broader impacts. The focus of this research and training is to increase the number of US citizens, including members of under-represented groups, who are advancing their teaching and research in the emerging areas of information assurance and security. The interaction of distinguished experts from National Laboratories, the LSU National Center for Security Research and Training provides coordinated efforts in security research and training by both state and federal law enforcement agencies. The integration of research and education, and broader cross-disciplinary activities is an important component of the inter-institutional collaborative work. This project expands the learning community and provides a way to enhance faculty's academic preparation and to establish partnerships for research and education in information assurance and computer security.

In this proposal, we propose to develop a Cyberinfrastructure for Reconfigurable Optical Networks (CRON). With supercomputing facilities and storage resources geographically dispersed across the US and around the globe, the success of scientific applications and collaborations relies increasingly on high-speed optical networks. The development and deployment of national and international optical networks and regional networks make it possible for more scientists, research and educational institutes to connect and collaborate at previously unachievable levels.

However, there is a critical gap between the research projects and the deployed high speed networks, caused by the following problems: (1) Limited physical resource ? e.g., the expensive resource of the LONI network has limitations on the number of simultaneous users, and there are many research and/or educational institutions in Louisiana that cannot access the networks due to the last mile problem.; (2) Unmodifiable environment - because the currently deployed networks are used for production as well as research, individual users cannot impose on other users by modifying environmental parameters, such as network protocol stacks and operating systems; and (3) Partial environment ? applications or protocols developed for one specific high speed network should work correctly over all kinds of different networks, irrespective of different network bandwidths and delays. However, each deployed physical network has one physical characteristic, such as bandwidth and delay.

The goal of this project is to bridge the gap between physical networks and research by developing a virtual networking and computing cyberinfrastructure, CRON, that provides integrated and automated access to diverse networking environments. To provide the virtual environments, CRON consists of two components: hardwares including a switch, optical fibers, network emulators and softwares automatically reconfiguring the hardwares based on users? demands. CRON resolves the critical issues by achieving three key objectives: (i) Unlimited virtual networking resource CRON will enable researchers to explore new network technologies and rapidly assess their impact on applications irrespective of physical limitation and allow educators to introduce the state-of-art networking environment to students who cannot access those physical networks. (ii) Reconfigurability CRON will allow researchers and educators to modify the virtual environments without interfering with others who share CRON. (iii) Complete environment CRON will provide all kinds of networking environments, including regional networks and global transoceanic optical networks.

Detection, identification, and tracking of CBRNE (Chemical, Biological, Radiological, Nuclear, and Explosive) plumes can be accomplished by combining the modalities of sensor and cyber networks. The sensor network provides information about physical-space activities, e.g., locations and movements of the plume sources. The cyber network provides storage and computational resources to analyze and infer where the plume originated, the trajectory of its movement, and the prediction of its future movement. The challenges in realizing such a sensor cyber network include intelligent sensing (intelligent sensor selection and coverage) and the capability to deal with uncertainties (uncertainty in measurement as well as in modeling). This project proposes to leverage the convergence between the sensor and the cyber networks to achieve these goals. In particular, the plan is to carry out three synergistic research tasks: 1) network formation by sensor selection, placement, and coverage; 2) sensor tasking protocol with temporal/spatial uncertainty management; 3) protocols for reliable sensor-cyber communication supporting the above two tasks. The PIs will prototype the research results and integrate the prototypes for different components to build the sensor cyber network for plume detection, identification and tracking. They will also evaluate the sensor cyber network using various test scenarios in collaboration with Oak Ridge National Lab. If successful, the project will provide technology for building detection and tracking sensor networks that can give great protection to people and the environment against harmful plumes.

The goal of this project is to construct an academic support program that meets the needs of financially disadvantaged students. Over the four-year grant period, 30 need-based scholarships are provided to qualified science, mathematics, and technology students at Louisiana State University. The students are participating in a variety of activities, including academic advising and mentoring, Individual Development Plans, academic and professional development activities, academic enhancement program, and research training activities provided by the faculty mentors in the sciences and math disciplines. The project increases the number of well-prepared undergraduates, including woman and minorities, entering graduate school or industry. The project also promotes inter-cultural participation among non-minority and minority students and faculty members.