Sriram Vishwanath, Ph.D. - US grants

Capacity of Wireless Networks: Cooperative Communication

The fundamental capacity limit of wireless networks is a highly pursued
yet elusive problem lying at the intersection of distinct fields of
research. The principle issue is the lack of appropriate tools within any
one field to tackle this problem in its entirety. This research activity
builds an inter-disciplinary framework for the analysis of this
fundamental limit, merging networking and optimization theoretic concepts
such as node cooperation, cooperative routing and Lagrangian duality in
conjunction with sophisticated mathematical tools from information theory.
A mathematical theory is developed using these diverse tools and, where
necessary, introducing new tools to obtain 1) improved achievable rates
for a given network configuration over those currently known 2) novel
outer bounds on network sum capacity, 3) non-trivial conditions where
inner and outer bounds meet and 4) algorithms that compute the
corresponding resource allocation strategies for time-varying wireless
networks.
Specifically, techniques such as joint source-channel coding, multi-user
binning and multiuser cooperative coding are used to develop the inner
bounds on the capacity of the systems, while Lagrangian duality theory and
estimation theory is used both to develop outer bounds on capacity and to
compare the inner and outer bounds.

Bacground

This project has been created with the aim of delivering the state-of-the-art research facilities in networking and communications the University of Texas at Austin to the vast underserved undergraduate population in the State of Texas. Students from seven distinct universities with programs in Engineering and Computer Science that form part the University of Texas System Arlington, Austin, Brownsville, Dallas, El Paso, Pan American and San Antonio will be actively recruited to participate in the EURECA endeavor. The program has been designed to ensure a significant participation of minorities and of disadvantaged communities in it. Minorities form more than 40% of the undergraduate population in the UT System, with UT Austin alone graduating the second largest number of minorities in the entire nation. In addition, UT Brownsville, El Paso, Pan American and San Antonio are minority institutions. Special partnerships have been formed between the seven universities to ensure the success of this program in reaching out and advertising, selecting worthy candidates and finally, in the follow-up process. Most importantly, our partnerships with other members of the UT System offer the knowledge and facilities possessed by UT Austins faculty to underserved students from institutions with limited resources and/or without graduate programs.

The Wireless Networking and Communications Group (WNCG) with thirteen professors and over five labs Network Engineering Lab, MIMO Communications Lab, Wireless Networking and RF Propagation Lab, Embedded Signal Processing Lab, Image and Video Engineering Lab and the new Optical Communications Lab forms the foundation for this project. All of the WNCG faculty members are part of the EURECA team, and will supervise undergraduate research as part of EURECA, while the WNCG staff will provide full logistical and administrative support for the endeavor. The EURECA team also includes educational research specialists to facilitate and improve this research interaction, a researcher from industry and one faculty member per University for all the other participating Universities.

The undergraduate research projects have been designed to provide participants with a gamut of options, with subgroups of faculty from WNCG spearheading each effort. Participants will be urged to interact with all of UT Austins EURECA members to identify their true calling in research, and will subsequently be assigned to a mentor that will guide them through their stay in Austin. A key component of every research project is its extensibility providing students with the capacity to build on knowledge provided to them by their mentor and the PI to generate independent results, and to sustain this effort after returning to their parent institutions. The pre-existing relationships between faculty in the UT System and the new partnerships formed will prove to be an invaluable resource that secures such an extensible research experience. In addition, the PI will deliver bi-weekly lectures to enhance the participants writing and oral presentation skills, and organize a mini-symposium to culminate the summer program.

The EURECA teams objective is to generate sustained research interest among undergraduates so that they become tomorrows leaders in industry and academia. This project has thus been meticulously designed to meet this goal.

Intellectual Merit
UT Austin is one of the premier research institutions in the nation, and the faculty members constituting the EURECA team have extensive experience both performing and conducting research. Many faculty members in this team also have considerable experience working with undergraduates from across institutions. The research projects in the EURECA endeavor have been created as a focused plan to guide the student from a passive learning state to the state of being an active self-sustaining researcher.

Broader Impact
Beyond impacting on the minority and underprivileged communities in Texas, and hence that of the entire nation, EURECAs research tasks also enhance the lives of the surrounding community. Projects such as one providing wireless access to impoverished sections of the local community proves to be an invaluable and satisfying experience to the undergraduate student, while inspiring youth in the community, specially women and other underrepresented groups, towards higher education in engineering.

ACLI-ware: Dynamic Data Driven Control for Wirelessly Implemented Application Systems

This project creates an end-to-end management solution for mobile computing that incorporates aspects from the physical wireless communication medium through application requirements. Our middleware
(ACLI-ware) moves abstracted information towards higher layers, making them channel aware, and conveys constraints and requirements towards lower layers, making them application aware. In providing applications channel awareness, information about the wireless channel, battery power, neighbors and their distances, lengths of routes, etc. are abstracted from the lower layers towards higher layers. For example, the appearance of a metallic wall may require selecting alternate routes and defining new communication clusters. On the other hand, a dense network requires clever interference-mitigation algorithms and fair-scheduling and routing algorithms. This information is also used by the application to throttle sending rates or adapt data fidelity. To achieve application awareness, applications' constraints such as latency, fairness, quality of service, etc. influence the algorithms used at lower layers. For example, tight latency constraints can lead to smaller communication clusters, shorter allowable routes and quick feedback algorithms from the physical medium. On the flip side, a high bandwidth/throughput requirement entails optimal clustering and high fidelity channel-feedback.

This project couples two components seldom brought together: and use of physical channel information for adaptation and the elicitation of application constraints to impact communication characteristics. Thus, there are two key innovations in ACLI-ware: (a) interaction amongst layers, with each layer depending on the other layers' inputs and (b) dynamic solutions, with algorithms that adapt in real-time to varying channel conditions and system requirements.

This project aims at developing a formal methodology for analyzing channel-aware random access in wireless networks. The unique features of this effort include: the development of models that incorporate, amongst other parameters, channel state as a central aspect of their formulation, and the use of cross-cutting tools to analyze the performance of such networks. The mathematical backbone from which the model and resulting solutions arise in this project are optimization and game theory. The specific results that will emerge during the course of this effort include: models that merge channel awareness and capture, a better understanding of the use of opportunism in random access, and finally, an understanding of the ways in which degrees of freedom can be allocated in multiple-antenna random access.

The impact of the underlying physical medium on the performance of access schemes can be considerable, particularly in wireless settings where the channel is highly dynamic and is impacted by myriad changes in the surrounding environment. Thus, a succinct representation of the physical medium combined with sophisticated analytical techniques to tackle the resulting models are essential to tap into the gains resulting from the use of channel-awareness in random access.

CIF: Medium: Collaborative Research: Cooperative Networking Across the Layers

This research goes beyond the physical layer in defining and analyzing cooperative techniques for wireless networks. By incorporating higher layer properties such as traffic dynamics and access control, the investigators develop a new theoretical framework for analyzing and designing cooperative networking algorithms across the layers, which includes existing cooperative techniques such as cooperative relaying and network coding. The basis of this research rests on two major points:: first, the realization that cooperative communication at the physical layer cannot be viewed in isolation, since it has implications at the access and network layers, and second, the recognition . that cooperation at the higher layers, in its own right, can significantly impact overall network performance.

This project has three inter-related thrusts: The first thrust studies resource allocation policies which stabilize the queues within various classes of cooperative networks, if stability is indeed attainable. The second thrust determines a family of scheduling algorithms which maximize the volume of traffic served by a cooperative network within a finite horizon. Finally, game theoretic models are developed for cooperative networks where nodes in the network are allowed to pursue differing objectives, and come to a distributed agreement on the (locally) optimal operating point for the overall network.
This research also considers non-stationary and non-ergodic environments that are more appropriate representations of the wireless channel in a network.

This site is co-funded by the Department of Defense in partnership with the NSF REU program. The site is also co-funded by the Networking Technology and Systems (NeTS) cluster of the NSF/CISE Division of Computer and Network Systems.

This award renews an exemplary existing Research Experience for Undergraduates site focused on wireless networking and communications. The project engages students in a wide range of projects in wireless networking lead by a team of faculty mentors in the Wireless Networking and Communications Group, including 5 labs. A key component of every research project is its extensibility providing students with the capacity to build on knowledge provided to them by their mentor and to generate independent results, and to sustain this effort after returning to students' parent institutions. Through a partnership with the Austin Technology Incubator the students will learn about the synergies between research and commercialization. A nationwide recruitment process is used to select cohorts of undergraduate students to participate in a nine-week summer research program at the host institution. Particular emphasis is placed on recruitment of students from under-represented populations. The project includes technical seminars and workshops, student presentations, and field trips and other professional development opportunities. The EURECA team's objective is to generate sustained research interest among undergraduates so that they become tomorrow's leaders in industry and academia.

Intellectual Merit: The intellectual merit of this project lies in strong research basis and the expertise of the faculty. The projects are in current research areas that are of interest to the community at large and that have clear practical applications. The research projects are created as a focused plan to guide the students from a passive learning state to the state of being active self-sustaining researchers.

Broader Impact: The broader impacts of the project include providing a quality research experience to undergraduate students, particularly students from underrepresented groups. The participating faculty members are committed to including under-represented minority students in their research, particularly through their partnerships with Huston-Tillotson University and professional organizations. Thus this project has the potential to produce new computer science graduate students and faculty members and to advance discovery and understanding while promoting learning.

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).

This project determines the fundamental limits of network secrecy from a network coding perspective, and then applies this theory to improve security guarantees in peer-to-peer and wireless networks. As network coding gains prominence as an important strategy for both wired and wireless networks, the project identifies both the advantages and vulnerabilities from using network coding. Subsequently, the effort develops a design methodology that exploits the advantages while carefully compensating for the vulnerabilities.

This project analyzes networks under both outsider and insider attacks. Specifically, coding mechanisms are developed to combat an external eavesdropper. Also, a combination of cryptographic and information-theoretic tools are used to combat internal modification attacks on the network. The results are then used in two case studies: eavesdropper attacks on wireless mesh networks and pollution attacks on P2P content distribution systems.

Secure network coded systems, once well understood, can greatly impact how networks are designed and deployed. Nearly every network setting (wireless, wired or heterogeneous) can benefit in terms of improved resilience (in addition to other performance benefits such as throughput) in its design. Case studies in this effort are designed to help transition the theoretical principles developed into practical algorithms.

The research team includes an industry member which will aid in transitioning our research ideas from theory to practice. The team will disseminate its findings through traditional scholarly venues, through the web and to the local community at each partner institution.

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).

The objective of this project is to inspire students at Tuskegee University and Auburn University in mathematics, aerospace science engineering, and networking by inviting them to contribute to a grand project: fly safely and efficiently, in a limited space, a fleet of autonomous UAVs on a cooperative mission with terrestrial vehicles.

The team of Dr. Sriram Vishwanath at the University of Texas (UT) Austin will deliver six Proteus modules (terrestrial vehicles) and will assist the team at Tuskegee University (TU) and Auburn University (AU) to build for this project six other Proteus modules adapted to the ultimate objective of this project. Dr. Biaz and his students will develop and implement at Auburn University the networking protocols that will allow all terrestrial and aerial vehicles to communicate with each other. The implementation will be made through series of laboratory exercises and research projects collaboratively conducted at Tuskegee University and Auburn University.

Intellectual Merit: This project will set up a research infrastructure that enables laboratory exercises in mathematics, aerospace and computer networking. For each discipline, students will contribute to the solution of challenging problems. To this day, most research on mobile ad hoc networks is conducted through simulation. This research infrastructure will allow students to test and evaluate the most promising networking protocols with hardware-in-the-loop. In mathematics, students will realize for example that solving a system of equations not only can obviously determine the position of a vehicle but also can be directly used to enable networking coding and improve communications efficiency.

Broader Impacts:
Tuskegee University is the only HBCU with an ABET-accredited Aerospace Engineering program and graduates the largest number of African-American aerospace engineers in the US. This research infrastructure will contribute to attract and retain students in STEM fields at Auburn University and Tuskegee University in general and under-represented minorities in particular. The multidisciplinary collaboration between Tuskegee University, UT Austin and Auburn University will benefit students from these institutions to work in a multi-disciplinary environment and will expose them to state of the art education and research. The results of the collaborative effort will be disseminated through presentations at engineering education and other professional conferences.

Abstract
0916713 - CIF: Small: Structured Transmission Strategies for Wireless Networks

In this project, we use lattice and other structured coding techniques to induce alignment in wireless networks. This effort uses these codes on three different fronts: a.) Interference networks, where we use structured codes to align the interference seen at each receiver. The objective is to determine the capacity of this channel to within a constant gap using these codes. b.) Cognitive networks - where we use lattice codes to mitigate the interference seen by both the licensed and the cognitive radios. We exploit code structure to both (partially) learn the interfering signal at the cognitive radio and then use this knowledge to precode/align our interference signal. c.) Secure wireless networks: again, we utilize the structure of the codebook to determine simple transformations at the source in order to keep eavesdroppers in the network at bay. We employ these codes to detect, and depending on code structure, correct for modification attacks on the codebook.

This project develops, from the ground up, a new theoretical framework for analyzing and designing algorithms for dynamic ad-hoc wireless networks. This proposal embraces network dynamics as an opportunity to be exploited, not an adversity to be overcome.

The approach is based on four inter-related thrusts:
1. Incremental Topology Learning: Tracking changes in the network much more efficiently than re-learning entire topology, using sparse "error graph" representations.
2. Topology and Traffic Shaping: Controlling the "effective" wireless network topology so that (i) at any instant of time it appears to be highly disconnected to scheduling algorithms, but retains global connectivity over time; and (ii) modifying traffic statistics to ensure statistical spatial correlation decay.
3. Warm-starting Distributed Algorithms: Message-passing algorithms that can warm-start the optimization based on local knowledge of past solutions.
4. Proteus - A Mobile Robot Testbed: This project validates its approach via implementation on a mobile robot testbed called Proteus, which is used to optimize algorithms in a practical setting.

Broader Impact: Industry is involved in this research from the start, via the WNCG Affiliates program at UT. The research will be disseminated via publications in top-tier venues, industry interactions, and specially organized workshops. Both graduate students and undergraduate students, via a REU program at UT (with emphasis on recruiting women and minorities), get exposure to both real-world wireless networks (via the testbed), and cutting edge theory.

We seek to investigate the application of advanced signaling and processing concepts, which have been developed for wireless communication, to multimode fiber-optic transmission systems. Traditionally, the maximum achievable data rate is limited by the distance-bandwidth product of the fiber; however, multiple-input multiple output signaling enables significant increases in bandwidth, by trading the fundamental limitations of dispersion for a computational exercise that that is well-understood. Our preliminary experimental results exceed what is achievable with traditional modulation schemes by over 24-fold, with the potential for orders of magnitude in additional performance, particularly with regard to bit-error-rate. In this program, we seek to expand our experimental and theoretical understanding of this new field by (1) building a low-cost, scalable, testbed using off-the-shelf components and (2) developing a complete theoretical framework.
The potential for greatly enhancing the bandwidth of low-cost optical links is particularly compelling for data centers as they rely on multimode fiber links at the rack-to-rack and board-to-board level. A method to greatly increase the achievable bandwidths and reduce the energy-per-bit penalty would be truly enabling. There are also substantial pre-existing deployments of multimode fiber in local area networks across the nation. Due to higher component costs, it is cost-prohibitive in many cases to replace this multimode infrastructure with single-mode fiber networks, as a means to meet future bandwidth demands. Using our approach, the bandwidth of such systems can be upgraded without resorting to single-mode fibers. Additionally, future drive-by-light, fly-by-light, and optical shipboard systems could also be enabled with this approach; such systems are attractive because of the potential to greatly reduce size, weight, power, and sensitivity to electromagnetic interference, as compared with conventional electrical signaling.
We believe that this effort is ideally suited to an EAGER because (1) the goals can be achieved rapidly and at moderately low cost; (2) it is a dramatic departure from conventional fiber-optic communication approaches with many fundamental questions remaining, rendering it virtually impossible to obtain funding through conventional mechanisms; (3) the effort is highly interdisciplinary (e.g. PI Bank is a photonic device and materials engineer, while Co-PI Vishwanath is an information theorist); and (4) upon answering the aforementioned questions, this approach can readily transition to more conventional funding mechanisms.

Intellectual Merit. The intellectual merit lies in advancing the theoretical underpinnings and experimental findings associated with the application of wireless communication approaches to multimode fibers. To this end, we will develop a framework for modeling multiple-input multiple-output communication over multimode fiber. This analysis will merge tools from information theory and statistical signal processing with those from photonics. There is limited work at the intersection of these disciplines and our efforts will make significant inroads into developing a comprehensive framework for characterizing the fundamental limits of multimode fiber communications. We will leverage this theory to design novel devices that are ideally suited to harnessing the potential benefits of multiple-input multiple-output strategies. We will also construct a system testbed using off-the-shelf-components, to study how performance scales with the number of transmitters and receivers, providing valuable feedback to the theoretical analysis.

Broader Impact. At the conclusion of this effort, we expect to demonstrate truly enabling capabilities in optical fiber communications, with myriad potential new applications. In addition to our research findings, which will be disseminated through journal publications and conference talks, we also seek to engage students at the high school, undergraduate, and graduate levels on the concepts and implications of optical fiber communication systems, their underlying foundations, and the importance of rigorous validation and evaluation. To this end, the PIs will develop presentations and demonstrations to engage high school students and teachers across Texas, through mechanisms including UT-Austin?s Edison Lecture Series, Summer Nanoscience Academy, and UTeach. We will also develop new undergraduate research projects on optical communications, to enhance the offerings of the NSF-REU site EURECA, directed by co-PI Vishwanath. This program targets participants from underrepresented groups enrolled at universities across the state of Texas. Additionally, as part of the team?s commitment to the nation?s student community, lectures generated by the PIs will be converted into course modules and made available online through The University of Texas? World Lecture Hall website.

This Small Business Innovation Research (SBIR) Phase I project aims to build a low cost, real-time 3D camera. The research objectives are to prototype a portable, compact and low power camera capable of capturing 3D models of indoor scenes. This research will use techniques from computer vision, robotics and high dimensional statistics and signal processing in order to achieve its goals. The novelty of the research is in bringing together cutting-edge ideas from these diverse disciplines to solve real-world problems in 3D modeling. The anticipated technical results include a working prototype that enables real-time operation without any human intervention.

The broader impact/commercial potential of this project includes the architecture, engineering and construction (AEC) industry. An industry where mistakes can prove costly both in terms of money and lives, our research can enable real-time tracking of the construction process, enabling workers, architects and engineers to ensure that mistakes are detected and corrected promptly during construction, and
historical monuments are preserved over time as 3D models. Additionally, this technology can aid educational activities in classrooms, help with remote surveillance etc., with multitude of applications to impact people from all walks of life.

Hardware authentication is one of the critical needs in the emerging discipline of design for assurance and design for security. It is concerned with establishing the authenticity and provenance of Integrated Circuits (ICs) reliably and inexpensively at any point in a chip's life-time. Physical unclonable functions (PUFs) have significant promise as basic primitives for authentication since they can serve as intrinsically-generated hardware roots-of-trust within specific authentication protocols. PUFs are pseudo-random functions that exploit the randomness inherent in the IC manufacturing to generate random output strings.

The central challenge in realizing the potential of strong PUFs is their vulnerability to model-building attacks using machine learning (ML) methodologies. Despite the effort devoted to PUFs over the past decade, there is still a lack of a strong silicon PUF that is both reliable and ML-resilient. This project develops a strong PUF that exploits the physics of nanometer-scale CMOS to attain security properties that are superior to existing designs. This new PUF exploits the physics of scaled transistors, namely subthreshold operation and drain-induced barrier lowering, to realize continuous nonlinearity. The project also investigates effective algorithmic techniques, based on novel codes and lattice-constructions, to enhance the overall structure as well as the robustness of the resulting PUF output to enable high-capacity secret key generation. The strong PUFs developed in this project possess a very large input-output space, making them invaluable to many potential security applications.

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project is to empower individuals to make financially sound healthcare decisions; helping them navigate the otherwise convoluted and confusing healthcare cost landscape by enabling them to manage, plan, and better understand healthcare expenditures and budget for upcoming and future costs. This project takes a data-driven approach, developing sophisticated, innovative, and large-scale data mining and machine learning algorithms to automatically learn a plethora of cost patterns from over a 100 million healthcare records.

The proposed project will provide a user-friendly, engaging interface for individuals to manage and understand healthcare expenses for themselves and their families. By bringing together ideas in data mining, machine learning and natural language processing to enable our technology, we make fundamental progress in research and development in the field of healthcare informatics. The anticipated results are the development of an algorithmic suite that can be used to model and predict the nature of healthcare costs across regional boundaries and demographic groups in the United States.

This award renews an exemplary Research Experience for Undergraduates (REU) site focused on wireless networking and communications at the University of Texas - Austin. The REU Site engages undergraduate students from institutions across the nation in a wide range of projects in wireless networking lead by a large team of faculty mentors in the Wireless Networking and Communications Group. The projects are in current research areas that are of interest to the community at large and that have clear practical applications. A key component of every research project is its extensibility providing students with the capacity to build on knowledge provided to them by their mentor and to generate independent results, and to sustain this effort after returning to their parent institutions. This award is co-funded by the Networking Technology and Systems (NeTS) program.

The intellectual merit of this project lies in strong research basis and the expertise of the faculty. The research projects are created as a focused plan to guide the students from a passive learning state to the state of being active self-sustaining researchers. A nationwide recruitment process is used to select cohorts of undergraduate students to participate in a nine-week summer research program at the host institution. Particular emphasis is placed on recruitment of students from under-represented populations. The project includes technical seminars and workshops, student presentations, and field trips and other professional development opportunities. The goal is to generate sustained research interest among undergraduates so that they become tomorrow's leaders in industry and academia.

Cloud storage systems increasingly form the backbone of software services that underwrite everyday lives, serving a large array of businesses and forming an essential pillar of the economy. Given the sheer volume of interactions within these distributed systems, there arise a multitude of issues in building, maintaining and enhancing them. The most salient of these challenges are in the reliability, availability, consistency, confidentiality and privacy of data stored in these vast systems.

The proposed research is expected to have a significant impact on the manner in which cloud storage systems are designed and deployed. In these systems, storage node failures can have a significant impact on the efficiency of the overall system. This project enhances fault tolerant mechanisms to enable efficient recovery from failures, while augmenting the overall data availability and privacy offered by such systems. The research effort will advance the science of cloud computing by developing a new family of algorithms for distributed storage , and connect the advances to the significant industry needs in this topic. The research agenda will also be tightly integrated with education and outreach activities with direct involvement of underrepresented minorities, graduate, and undergraduate students.

Focusing on efficient maintenance of data with a range of desirable qualities, including mechanisms that ease data encoding, accessibility, updates, as well as privacy, the main objectives of this effort include (i) to develop coding schemes where a failed element can be regenerated with higher repair efficiencies from its local neighbors, (ii) to bring together the advantages of both local decodability and local repairability into one coding solution, (iii) to design mechanisms that provide low cost data updates in addition to efficient repair, (iv) to develop codes that can be resilient against failures with different scales/modalities, (v) to develop coding mechanisms taking advantage of implementation aspects of existing systems, and (vi) to develop coding schemes that enable users to access their data in a private manner. This effort addresses these challenges using a combination of tools from disciplines spanning coding theory, information theory, communications, as well as combinatorial and discrete mathematics.

The goal of this project is to enable interference alignment in practice using the concept of reconfigurable antennas, and by doing so, enable a much deeper understanding of the "interference barrier" that plagues wireless systems today. Interference alignment is a new, novel transmission strategy that results in linear gain in throughput with the number of nodes in the network. The planned research is expected to have a significant impact on the manner in which next generation wireless systems operate and how they access spectrum in multiple domains, including public safety, inter-vehicular communication and Industrial Internet of Things (IoT). The research agenda will also be tightly integrated with education and outreach activities at University of California-Irvine and University of Texas-Austin, with direct involvement of underrepresented minorities, graduate, and undergraduate students as well as a minority focused Research Experiences for Undergraduates program at University of Texas.

The effort initially begins with a demonstration that blind interference alignment (with no channel state knowledge at the transmitter) is indeed implementable in hardware, in real-time . Moreover, the gains promised by interference alignment in theory are indeed realized in practice for a 2 by 2 X channel setup. Using this as the foundation, it goes on to propose to design, characterize and develop reconfigurable-antenna-based interference alignment solutions in a variety of more general settings, including those where we have partial/delayed/mixed/distributed and/or alternating channel state knowledge. The effort is complemented
with implementations on a variety of platforms. The primary goal is to improve spectral efficiency of wireless networks.