|
Home >Research > Research Areas > Systems Research Projects
|
|
| Systems Research Projects
|
|
|
Nonlinearity and Uncertainty in Control System Design Members: R. Freeman Sponsor: National Science Foundation (CAREER Award) A basic purpose of feedback is to reduce the effect of uncertainty on the behavior of a system. When applied inappropriately, however, feedback can also increase the risk of instability and other undesirable phenomena. An obstacle to the successful design of feedback controllers has long been the presence of nonlinearity in models and in real systems. The goal of this project is to develop innovative methods for designing feedback controllers for uncertain nonlinear systems and to educate future engineers and researchers in their application. The research will identify flexibilities in new and existing nonlinear design methods which can be exploited to obtain good controllers for practical applications.
A behavioral approach to dissipativity analysis in nonlinear systems, with applications to human/robot interfaces Members: R. Freeman Sponsor: National Science Foundation This project will focus on the investigation of some of the basic unresolved issues in the study of the performance properties of nonlinear systems. Once these issues have been addressed, the resulting theory will be applied to the stability analysis of various systems in which single or multiple human operators are in feedback loops with active nonlinear robotic devices and are thus subject to applied forces.
Adaptive Reduced-Rank Interference Suppression Members: M. Honig Sponsor: National Science Foundation This project explores algorithmic, performance, and hardware issues related to reduced-rank adaptive filtering. Such filters project a received signal onto a lower-dimensional subspace to reduce the amount of training data needed relative to conventional full-rank algorithms. A goal is to build a low-power, special-purpose, hardware prototype which can serve as the computational engine for reduced-rank filtering in a variety of applications. Algorithmic issues include selection of filter rank, performance in various adaptive-filtering applications, numerical stability, and dynamic range.
Adaptive Signal-Processing Techniques for Spread-Spectrum Multiple Access Members: M. Honig Sponsor: Army Research Office This project is concerned with signal processing and coding techniques for Direct-Sequence (DS)-Code-Division Multiple Access (CDMA) in a peer-to-peer network. Topics to be investigated include the performance of adaptive interference suppression algorithms in the presence of wireless channel impairments, combined coding and adaptive interference suppression, and joint adaptation of transmitters and receivers.
Dynamic Resource Allocation for Code-Division Multiple Access Members: M. Honig Sponsor: National Science Foundation This project is concerned with techniques for supporting integrated services over a wireless cellular network. A fundamental problem is how to allocate radio resources to different users to optimize an overall network objective criterion while satisfying quality-of-service constraints. Wireless networks based on Direct-Sequence (DS) – Code-Division Multiple Access (CDMA) networks are considered, since DS-CDMA is a strong contender for next-generation cellular systems. Radio “resources” in a DS-CDMA network include the transmitted powers, processing gains, and number of assigned codes. The focus of this research is on dynamic resource allocation in the presence of time-varying user requests. The project’s goals are: (1) develop techniques for dynamic resource allocation given multiple traffic flows with different quality-of-service requirements; and (2) characterize performance / complexity tradeoffs associated with these techniques given simple traffic and propagation models.
Multi-Cell Dynamic Resource Allocation Members: M. Honig Sponsor: Motorola This project studies utility-based resource allocation methods for wireless networks. Each user is assumed to have a utility function which indicates the utility derived from the service vs. the amount of resource (e.g., power, time slots, or codes) allocated to that user. A typical objective is to allocate resources to maximize total utility or revenue summed over all users. Pricing strategies are being investigated which may accomplish this objective. Both single-cell and multi-cell models with other-cell interference and mixed services (voice and data) are being considered.
Audio-Visual Synthesizer for Telephone Communications Members: A. Katsaggelos Sponsor: Motorola Center for Communications This project develops a speech driven visual communication device. Speech is analyzed and used to synthesize visual speech features. By efficiently utilizing the correlation between audio and visual information, speech intelligibility can increase for impaired hearing people or for people who use telephone communication systems in noisy environments.
Development of Undergraduate and Graduate Programs in Medical Imaging Members: A. Katsaggelos Sponsor: Whitaker Foundation This project is aimed at creating an educational program which includes: 1) new coursework in imaging integrated into the biomedical engineering (BME) curriculum; 2) new research possibilities for BME graduate students; 3) experiences with the use of imaging techniques in clinical medicine; and 4) experiences in industry. This is an extensive, formalized program which takes advantage of the unique capabilities in basic, clinical, and industrial research at Northwestern University.
Digital Image Restoration Members: A. Katsaggelos Sponsor: Hitachi This project deals with the development of new digital image restoration and enhancement techniques applicable to scanning electron microscope (SEM) images. The focus of the research effort is on the development of learning-based algorithms for blur identification and image restoration. Such algorithms deviate from the traditional approaches in this area by utilizing priors that are learned from similar images instead of using generic priors.
Image and Video Processing and Communications in the SenseIT Program Members: A. Katsaggelos, T. Pappas, and R. Berry Sponsor: Defense Advanced Research Projects Agency A main objective in the Sensor Information Technology (SensIT) program is to provide robust detection and tactical intelligence using distributed-sensor signal processing and network-based intelligent agents. Within this program, we are determining: 1) the role of imaging and video over a distributed set of low-power wireless nodes within a tactical scenario; 2) the appropriate distributed architectures; 3) the imager and node capabilities; 4) the power, image resolution, and video temporal resolution (frames per second) requirements for a tactical scenario; 5) novel network architectures for the transport of images and video over multi-hop wireless networks; 6) energy-efficient protocols taking into account both the energy required for processing as well as transmission of images; and 7) the use of error resilience and concealment techniques to compensate for lost or corrupted packets, taking into account the high loss rate and unreliable nature of a multi-hop wireless network.
Immersive Virtual Reality Members: A. Katsaggelos Sponsor: Motorola Center for Communications The major goal of this research is to develop new techniques for creating a complete virtual environment. The ultimate immersive virtual environment should provide three elements: (1) stereo vision, where each eye receives a different image according to its location in space; (2) complete 360Ú (panoramic) view, allowing the user to look in any desired location; and (3) video views in any direction. None of the currently known applications is able to provide all three elements simultaneously. A set up is under development for capturing multiple video sequences and generating stereo video panoramic views at video rates. The transmission of such views is very bandwidth consuming, therefore requiring a compact representation of stereo panoramic data. Development of appropriate compression algorithms is also an important goal of this project.
Pre- and Post-Processing Techniques for Video Compression Members: A. Katsaggelos and T. Pappas Sponsor: Motorola Center for Communications In this project pre- and post-processing algorithms are developed for increasing the coding efficiency of a coder and the visual quality of the reconstructed video. The algorithms under development are standard compliant, finding therefore wide applicability. We are primarily concerned with high bit-rate video compression applications (up to 10 Mbps) using the MPEG-2 and MPEG-4 standards and CCIR601 resolution video. We are also primarily concerned with real-time (video rate) implementation of such algorithms.
Three-Dimensional Motion Estimation Members: A. Katsaggelos and R. Lagendijk (Delft University of Technology) Sponsor: North American Treaty Organization The project deals with the estimation of the motion and disparity fields in stereoscopic video sequences. By exploiting the interrelation between these two fields, better estimation results can be obtained. The resulting disparity field can be used, for example, for the generation of intermediate views in binocular video sequences.
Video Compression Members: A. Katsaggelos Sponsor: Northwestern University Research Administration Award This project deals with the efficient encoding of object boundaries in the rate-distortion sense, when the available bit budget (storage applications) or bandwidth (transmission applications) is limited. This problem is becoming increasingly important in multimedia applications such as object-oriented video coding, content-based storage and retrieval, studio and television post-production and mobile multimedia applications. The newly adopted video compression standard, MPEG-4, allows for the transmission of arbitrarily shaped video objects. The video scene is decomposed into video objects, each of which is represented by its shape and texture and can be manipulated independently. Efficient and rate-distortion optimal techniques have been developed for approximating a curve with polynomials of any order.
A Control Theoretical Approach to Wireless Multimedia QoS Members: C. C. Lee and A. Haddad Sponsor: Motorola An analytical framework is very important for comprehensive network bandwidth management in a differentiated quality-of-service environment. Our approach includes an adaptive traffic source model that accounts for a wide variety of network traffic types. Based on this model, we designed a systematic means for managing throughput, delay, and loss on both a per-flow basis and an aggregated basis. These performance variables are readily translated to various quality-of-service attributes and/or utility functions of interest. Accordingly, we designed an analytical core for designing integrated call admission, bandwidth allocation, and rate-adaptation algorithms for provisioning differential quality of service.
Novel Self-Organizing Wireless Network Protocol Members: C. C. Lee Sponsor: Motorola We are employing analyses and simulations to design multiple-layer wireless network architectures and protocols for reliable and efficient multiple-hop communications between low-cost, low-power, self-organizing nodes. Each layer uses a high-power root node to supervise the self-organizing functions, to coordinate medium access control, to maintain the physical topology, and to serve as the root of the hierarchical routing backbone of the layer. In one research area, we are designing low-complexity root-positioning and repositioning algorithms that optimize two main network performance metrics: (1) a topological optimization that provides a shortest path for every node to reach a root; and (2) a minimization of the overall network traffic for a given layer scheduling algorithm. In a second research area, we are developing low-complexity layer-scheduling algorithms and associated control protocols that adapt to network load fluctuations, thereby optimizing the performance on a continuing basis. While one type of performance metric aims at load-balancing the layers such that the performance is optimized on a global scale, the other type strives to optimize the delay of individual message transmissions.
Instrumentation to Correlate Body Position with Cardiac Arrhythmias in Ambulatory Patients Members: A. Sahakian and S. Swiryn (Evanston Hospital) Sponsor: O’Shaunessy Foundation Some patients have cardiac arrhythmias or experience dizziness or loss of consciousness caused by physical activities. It is often quite difficult to diagnose and manage such patients. This project is developing the technology to simultaneously monitor and interpret both body activity and the electrocardiogram using a small patient-worn device.
Lightwave Cryptographic Techniques Members: H. Yuen and A. Sahakian Sponsor: Defense Advanced Research Projects Agency The objective of this project is to develop new cryptographic techniques, and to modify the important existing ones, for applications to encryption and authentication in energy-constrained sensors with limited memory and computational capability. The goal is to minimize power consumption in order to maximize the lifetime of the sensor operation and the amount of useful processing that can be carried out within the lifetime.
Mechanisms of Atrial Fibrillation, a Study Focusing on Issues of Repolarization Members: A. Sahakian and S. Swiryn (Evanston Hospital) Sponsor: Dr. Scholl Foundation Atrial fibrillation is the most common abnormal heart rhythm, affecting millions of Americans with symptoms such as palpitations or difficulty in breathing, and increasing the risk of stroke. Though there are palliative treatments available, there is only limited understanding of the mechanism. Perhaps because of this, there is only limited success in treatment. Much research has focused on depolarization (the electrical activation of heart cells) and its patterns during fibrillation. In this project, we are studying repolarization (the recovery of cells after activation). This may be a more important factor in the development, maintenance, and termination of atrial fibrillation.
Pacemaker Education Projects Members: A. Sahakian Sponsor: Medtronic, Inc. Cardiac pacemakers are designed to last as long as ten years and must therefore use as little power as possible; a typical total current drain is 30 microamperes. Achieving this while also guaranteeing reliable pacing requires innovative circuit designs and careful selection of pacing pulse parameters. This project has as its goal the development of mathematical simulations and prototype circuits which model the effects of pacing pulse variables on power consumption.
Signal Processing for Implantable Cardiac Pacemakers Members: A. Sahakian Sponsor: Medtronic, Inc. Implantable cardiac pacemakers have evolved from simple free-running oscillators into sophisticated devices which can diagnose and treat problems as they arise. New signal-processing algorithms are necessary to permit pacemakers to make reliable diagnoses. We are working toward highly reliable, low-power-consumption algorithms to diagnose selected classes of cardiac arrhythmias.
Vanderbilt-Northwestern-Texas-Harvard/MIT Engineering Research Center in Bioengineering Educational Technology Members: A. Sahakian and 13 others at Northwestern Sponsor: National Science Foundation This multi-university Center is working toward new developments in the biomedical engineering curriculum. Currently, A. Sahakian is developing new curriculum components for classes teaching state-of-the-art applications of signal processing in medical devices. |
|
|
Copyright
© 2002 Department of Electrical and Computer Engineering, |
|