Select a calendar:
Filter July Events by Event Type:
SUNMONTUEWEDTHUFRISAT
Conferences, Lectures, & Seminars
Events for July
-
Wireless Channel Uncertainty in Relay-Assisted Communication and Distributed Detection Systems
Mon, Jul 18, 2011 @ 11:00 AM - 12:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Azadeh Vosoughi, University of Rochester
Talk Title: Wireless Channel Uncertainty in Relay-Assisted Communication and Distributed Detection Systems
Abstract: One of the main challenges in wireless communications is coping with channel uncertainty. Dealing with this uncertainty, and the limitations it imposes, is tightly related to the specific system and its application. In this talk, we consider two systems, namely a wireless bi-directional relay-assisted communication system and a wireless distributed detection system. We study the impacts of channel uncertainty on the performance limits of these two systems and investigate optimal transceiver designs that minimize these impacts.
For the bi-directional relay-assisted communications we consider a training-based system, in which receivers learn the channels via employing dedicated pilot symbols. Assuming Gaussian inputs and block Rayleigh fading channel model, we study the trade-off between the accuracy and the bandwidth/energy costs of channel estimation and explore optimal transmit resource allocation, subject to network power constraint. We consider Cramer-Rao lower bound for channel estimation, sum-rate and outage probability bounds as the performance metrics.
Next, we discuss the effects of channel uncertainty on the design and performance of a wireless distributed detection system that is tasked with solving a binary hypothesis testing problem. We consider systems with training-based and blind channel estimation and coherent/non-coherent receptions. We investigate the optimal data fusion rules that maximize the overall system detection reliability and error exponent. Furthermore, we present and compare several detection and data fusion designs that exploit diversity to combat channel uncertainty and enhance system performance.
Biography: Azadeh Vosoughi is Wilmot Assistant Professor in the Department of Electrical and Computer Engineering at the University of Rochester. She received her BS degree from Sharif University of Technology, Tehran, Iran, in 1997, her MS degree from Worcester Polytechnic Institute, Worcester, MA, in 2001, and her PhD degree from Cornell University, Ithaca, NY, in 2006, all in Electrical Engineering. Her research interests lie in the areas of wireless relay-assisted communications, distributed detection and estimation, and distributed source coding and compression. She was the recipient of the Furth award in 2006 and was appointed as Wilmot Assistant Professor in 2009 at the University of Rochester. Dr. Vosoughi received the NSF CAREER award in 2011 for her research on the integration of signal processing and communications for distributed detection systems.
Host: Urbashi Mitra, ubli@usc.edu, EEB 536, x04667
More Information: Vosoughi.pdf
Location: Hughes Aircraft Electrical Engineering Center (EEB) - 248
Audiences: Everyone Is Invited
Contact: Gerrielyn Ramos
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor. -
Pd.D. Defense
Tue, Jul 19, 2011 @ 02:30 PM - 04:30 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Yi-Hua Edward Yang, USC EE-Systems
Talk Title: Large-Scale and High-Throughput Pattern Matching on Parallel Architectures
Abstract: Large-scale pattern matching has many applications ranging from text processing to deep packet inspection (DPI) where hundreds or thousands of pre-defined strings or regular expressions (regexes) are matched concurrently and continuously against high-bandwidth input data. The large number of patterns and the high matching throughput make large-scale pattern matching both compute and memory intensive. In this thesis, we propose novel algorithms, constructions, and optimizations to accelerate large-scale pattern matching on two prominent classes of parallel architectures: Field Programmable Gate Arrays (FPGA) and general-purpose multi-core processors. We focus our studies on large-scale string pattern matching (SPM) and regular expression matching (REM) in the context of DPI for network intrusion detection.
For SPM, we propose a head-body partitioning to efficiently handle large dictionary. Specifically, we design a pipelined affix-search relay for accelerating the dictionary "head" on FPGA, as well as a body branch data structure with branch grafting for accelerating the dictionary "body" on processor core with single-instruction multiple-data (SIMD) operations. For REM, we propose a modified McNaughton-Yamada algorithm to convert an arbitrary regex into a modular NFA with uniform structure, which can be (1) mapped onto FPGA with spatial stacking and parallel SRL for scalable throughput and efficient resource utilization; (2) further partitioned into a segmented regex-NFA for efficient implementation in word-based operations on general-purpose processor core. We also develop software framework to automatically optimize our REM solutions. Overall, our designs achieve better memory efficiency, higher per-stream throughput, faster construction and better attack resilience over previous state-of-the-art solutions.
Finally, we formalize a novel semi-deterministic finite automaton for REM, offering space-time tradeoff between the compute-intensive NFA and the memory-intensive DFA. We propose the convolvement analysis and compatible state grouping algorithms to convert any NFA into a minimal SFA, whose space efficiency can then be traded off for lower time complexity.
Biography: Yi-Hua Edward Yang received the B.Sc. degree in Electrical Engineering from National Taiwan University, Taipei, Taiwan in 1997, and the M.Sc. degree in Electrical & Computer Engineering from University of Maryland, College Park, Maryland in 1999. He was a research assistant at the Information Sciences Institute, University of Southern California From 2005 to 2007. He is currently a Ph.D. candidate in the Ming Hsieh Department of Electrical Engineering, University of Southern California. His research interests include algorithmic optimization on parallel architectures, high performance architecture designs for network processing, security and cryptography.
Host: Defense Chair, Prof. Viktor K. Prasanna
Location: Hughes Aircraft Electrical Engineering Center (EEB) - 248
Audiences: EES & CS PhD Students & Faculty
Contact: Janice Thompson
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor. -
Large-Scale Industrial Software Systems: Research Opportunities and Challenges
Fri, Jul 22, 2011 @ 11:00 AM - 12:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Dr. Srini Ramaswamy, Industrial Software Systems, ABB Corporate Research Center, India
Talk Title: Large-Scale Industrial Software Systems: Research Opportunities and Challenges
Abstract: Software systems development is fast becoming a globalized activity and this is an increasingly major trend within all industrial sectors. Due to the many benefits of globalization, from the integration of multiple ethnic / market perspectives driven idea generation to development cost structuring, middle and small-sized software companies are now beginning to establish worldwide development campuses / partners. Thus, globalization has become an overwhelming phenomenon in the software industry and is rapidly defining the nature of software development for the 21st century. For Industrial Automation companies like ABB in emerging markets such as India, these opportunities are both exciting we well as immensely challenging. They present problems that are incredibly different from similar-sized western markets and require a significant amount of innovation and creativity to develop robust, sustainable, yet significantly low-cost solutions for such markets. In this talk, I will present an overview of ABB in India and its research activities, specifically in the areas of Industrial Communications and Industrial Software Systems.
Biography: Dr. Srini Ramaswamy transitioned from an academic to a corporate research career in 2010, as the head for Industrial Software Systems research at ABB India Corporate Research Center, in Bangalore, India. His primary role is in research team building and leadership, developing university relationships and engaging in applied research for the creation and execution of projects with transformative value for the company's power technologies and process automation business units. On the academic front, he also serves as a visiting professor at the University of Arkansas at Little Rock and a honorary adjunct professor at the Indian Institute of Information Technology â Bangalore. His research interests are on intelligent and flexible control, behavior modeling, analysis and simulation, software stability and scalability; particularly in the design and development of complex software systems. Specific applications include real-time control issues in automation and manufacturing, data mining and distributed real-time applications. His work is motivated by the desire to understand the various requirements to build scalable, intelligent software systems with the inherent ability to successfully respond to observed and reported behavioral changes in their environment.
Host: Vice Dean Raghu Raghavendra
Location: Ronald Tutor Hall of Engineering (RTH) - 324
Audiences: Everyone Is Invited
Contact: Janice Thompson
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor.
