The ASU Signal Processing and Communications (SPCom) Group
Department of Elecrical Engineering, ASU

August 2006

GWC-4th Floor

Members:
Chaitali Chakrabarti, Professor, Low Power Architectures
Douglas Cochran, Associate Professor, Statistical Signal Processing, Sensing
Tolga Duman, Associate Professor, Wireless Communications, Channel Coding
Qian Gang, Assistant Professor, Face/Gait Recognition
Joseph Hui, Professor, Networks, Optical Networks
Lina Karam, Associate Professor, Image/Video Processing and Coding
Darryl Morrell, Associate Professor, Detection and Estimation, Sensing
Martin Reisslein, Associate Professor, Networks for streaming data and media
Andreas Spanias, Professor, Speech/Multimedia DSP, Adaptive Acoustic Sensing
Antonia Papandreou-Suppappola, Associate Professor, Sensing, Time-Frequency Analysis
Cihan Tepedelenlioglu, Assistant Professor, Signal Processing for Communications, Multicarrier systems
Harvey Thornburg, Assistant Professor, Signal Processing and Computer Music
Junshan Zhang, Associate Professor, Information Theory, Ad-Hoc Networks

News/Info (Aug 2006)

The ASU Signal Processing and Communications Group (SPCom) is part of the Department of Electrical Engineering. Research activities of the 13 faculty members in this group are supported by laboratory facilities representing the following focus areas: Digital Signal Processing, Sensor and Information Processing, Communications and Multimedia Networks, Image Processing, Java Systems and Speech and Audio Processing. Over the past seven years, five SPCom faculty members have received NSF CAREER awards. Electrical engineering graduate students are key contributors to the group’s research, and several SPCom doctoral graduates now hold prestigious faculty and research positions at the University of Texas, Brooklyn Poly, MIT Lincoln Laboratory and IBM Research. SPCom’s visibility has been enhanced by its working relationship with Raytheon on sensing applications, its collaboration with ASU’s Arts, Media and Engineering program and its partnership in NIH activities. Research sponsors of the group include DARPA, AFOSR, the Office of the Secretary of Defense, the NSF, General Dynamics, Motorola, Intel and Seagate. The group has founded and led the FSE Sensor, Signal and Information Processing (SenSIP) cluster, which is in the process of transition into a multidisciplinary research center with international activities.

 

Sensing and Information Processing
Sensors are ubiquitous in today’s technology products and systems. From power plants to medical devices, navigation to safety, sensors are increasingly important in many aspects of our daily lives. Spurred by advancing electronic and optical technologies, highly advanced, agile sensors are emerging as a next-generation technology for defense applications, such as surveillance, target tracking and structural health monitoring. Over the past few years, Professors Cochran, Papandreou-Suppappola and Morrell have been working to integrate the mathematical foundations of sensing and processing with a special emphasis on developing new algorithms to exploit the agility of emerging sensor systems. Towards this goal, they have received significant DoD funding from DARPA (). Members of the SPCom group are currently involved in two DoD Multi-Disciplinary University Research Initiatives (MURI) projects, which are providing $11 million in research funding to ASU over five years. The faculty’s research on these projects, one of which ASU is serving as the lead, entails collaboration with Raytheon, AFRL, NRL, Princeton, Purdue, Harvard, the Universities of Maryland and Melbourne. Other noteworthy collaborative activities in the sensing area include multimodal sensing with AME and analysis of ion-channel sensor signals (Professors Spanias, Goodnick and Thornton).

 

Speech and Audio Processing
The Speech and audio coding initiative is led by Professor Andreas Spanias and spans research in perceptual speech and audio coding and adaptive signal processing. It also incorporates several other research foci underpinning new technological capabilities, including adaptive signal enhancement, which utilizes vector quantization, Hidden Markov models (HMM) with extensions to denoising and classifying biomedical signals.

Sponsored work in speech processing includes a 5-year, $1.5 million dollar ASU program, which was funded by Intel Corp. Under this initiative, SPCom researchers contributed speech coding software for the Intel ProShare teleconferencing software package and to the design of a low-power Intel DSP core 60172. These contributions were recognized by awards from Intel Corporation citing technical leadership and outstanding contributions. Audio coding work was recognized by the IEEE Donald Fink award in 2002.

 

Image and Video Processing
The work on error-resilient image and video compression enables the transmission of multimedia information over low-bandwidth and error-prone channels such as satellite and other wireless channels. Professor Lina Karam’s work on image and video compression, enhancement and transmission, including region-of-interest (ROI) coding, has been integrated by General Dynamics into their SelectFocusTM Image and Video commercial products. Her research in this area has also successfully demonstrated, for the first time, the wireless transmission of digital imagery and video over the worldwide Iridium satellite communication system. As a result, digital multimedia data, can now be transmitted globally, even to remote and isolated areas with no special-purpose multimedia infrastructure. This is crucial for outreach efforts, disaster management and many applications including telemedicine, distance training, remote sensing and surveillance.

 

Communication and Networks

A prime goal of SPCom telecommunication research is to understand the traffic and quality statistics of encoded video and the resulting implications for multimedia transport over networks such as the Internet. By performing wireless networking research that examines efficient clustering, routing and media streaming in mobile ad hoc networks (MANETs), Professor Martin Reisslein has produced the first MANET routing protocol with complexity less than the total network size. This optical networking research develops novel access and metropolitan area networks that will provide bandwidths on the order of Gbps to end users. The focus of SPCom faculty has also been shaped by the continuously increasing demand for high-rate, reliable communication services across the world. ASU has received funding from NSF ITR in areas associated with wireless connectivity and ad hoc networks supported by antenna arrrays. Professor Cihan Tependelenliogu has centered his research on modeling the wireless communications channel, estimating its parameters, and analyzing and designing the modulation of coding schemes such as Orthogonal Frequency Division Multiplexing, ultra-wideband communications and adaptive modulation and coding for single and multiple antenna systems. Additionally, Professor Papandreou-Suppappola researches time-varying signal processing for wireless communications including time-frequency techniques for modulation and channel modeling. Research in cross-layer optimization and control in ad hoc sensor networks is being pursued by Professor Junshan Zhang and his efforts . have been recently recognized by a prestigious young investigator award from the Office of Naval Research (ONR). SPCom’s research has also been advanced by Professor Duman’s study of underwater acoustic (UWA) communication.. His research team is collaborating with Space and Naval Warfare Research Systems Center (SPAWARSYSCEN) and Heat, Light, Sound (HLS) Research, to address the UWA communication needs of the U.S. Navy The team is investigating the applicability of multiple-input, multiple-output (MIMO) for UWA channels. MIMO, is an exciting technology that may address such obstacles as multipath and fading in underwater communications. By using a combination of sophisticated channel coding, multi-carrier modulation and powerful iterative equalization techniques, Duman’s team has already increased the effective throughput of the shallow water links by close to an order of magnitude, demonstrated in actual at-sea experiments. Such improvements will open the frontiers for several different classes of applications, such as real-time image and video transfer and underwater networks, which were previously thought to be too demanding for practical implementation on UWA links. These UWA efforts have been funded by SPAWAR and the Navy’s small business transfer technology program, ONR STTR.

 

Low-power VLSI Signal Processing
Traditionally, research in design and implementation of signal processing systems has focused on finding the best way to map an already designed algorithm into an architectural platform. However, by modifying the specifics of the algorithm to suit the constraints of the architecture, Professor Chakrabarti believes that a more efficient implementation can be achieved. Her research team is collaborating with researchers at Duke University and Penn State University on a DARPA-funded project on an automated framework for algorithm-architecture co-design for FPGA platforms. This research team is also sharing NSF funding with the University of Michigan for a project involving an algorithm-compiler-architecture co-design strategy for designing an ultra low power baseband processor of a software-defined radio. Due to the increasing demand for portability, low-power systems are a priority for SPCom researchers. In order to design such systems, power has to be reduced at al levels of the design -- from algorithm level down to gate level. At the algorithm level, Chakrabarti’s team has been able to show how power reduction can be achieved by migrating seamlessly to a lower complexity algorithm during run-time in response to changes in channel conditions or quality requirements.

JAVA-DSP Development
Another achievement of SPCom is the development of a Java digital signal processing software package (J-DSP). Designed by Professor Spanias and his team of collaborators and graduate students, J-DSP was ranked by the Berkeley NEEDS committee as one of the top three educational resources in 2003. This software allows students to design and manipulate the basic building blocks of a DSP system and to experiment with simulations of digital cell phones, MP3 compression, and real time sensing. Spanias is building on the success of this software by spearheading an effort at ASU to develop a multi-disciplinary distance learning initiative that will train the next generation engineers. With the aid of $1.1 million in NSF grants, ASU is working with four other universities to enhance, evaluate, and disseminate the Java technology. Several Electrical Engineering faculty members, including Professors Duman, Karam, Papandreou-Suppappola, Tepedelenlioglu, Tsakalis, and Zhang, have collaborated with Spanias on some of these J-DSP efforts.

Arts, Media and Engineering
SPCom faculty collaborate formally with the Arts, Media and Engineering (AME) program and an established degree concentration is available for graduate studies in this area. Professor Spanias, associate director of AME, is Co-PI on a $3 million NSF IGERT grant that supports some of the AME research activities. Research, led by Electrical Engineering Professor Gang Qian entails examining movement in a holistic way and trying to teach computers to understand this language in much the same way that humans do. Such research can result in a more natural movement-based mechanism for human computer interaction.

Professors Thornburg and Spanias’ work with AME includes a source localization project with microphone arrays. While Spanias and his students have developed adaptive algorithms for microphone arrays, Thornburg is developing new methods, based on probabilistic models that use dynamic Bayesian networks, to segment, analyze and recognize patterns in human activity occurring in situated environments. Ongoing applications include joint gesture segmentation and temporal structure inference from conducting performance, and audio summarization of continuously-monitored everyday sound environments. AME work has been applied to biofeedback for rehabilitation, K-12 mediated education and interactive dance performance.

Biomedical Signal Processing Applications
Students and faculty working in the signal and image processing labs are actively participating in the development of next-generation techniques for magnetic resonance imaging (MRI), including methods for medical data collection and reconstruction and motion-corrected imaging. Such techniques can result in better and more accurate diagnosis and, consequently, better prevention and faster cure. This work is performed in collaboration with Dr. Jim Pipe in the MRI department at the Barrow Neurological Institute (BNI). Other collaborative efforts of SPCom faculty and students in the biomedical field include epileptic seizure prediction (with Professor Iasemides) and DNA sequence analysis using spectral estimation techniques.