Event Details

A Particle Filter Solution for Single Platform Passive Doppler Geolocation with Unknown Emitter Frequency Speaker: Hanna E. Witzgall, SAIC Fellow Abstract: This seminar discussed a novel application of particle filters to the problem of single platform, passive Doppler geolocation. Particle filters implement a recursive Bayesian filter by representing the posterior density with a set of discrete particles and their associated weights. They have been used to great success in a wide variety of problems. The seminar discussed particle filtering methods in general and then described a particular particle filtering solution to efficiently solve for the geolocation of a radio frequency emitter using passive Doppler-shifted frequency measurements. Specifically, the new technique addresses the issue of unknown emitter frequency for the Doppler geolocation problem. The Global Navigation Satellite System Speaker: Dr. Christopher J. Hegarty, ION Fellow Abstract: The Global Navigation Satellite System (GNSS) is the worldwide set of satellite navigation constellations, civil aviation augmentations, and user equipment. This presentation reviews the current status and future plans of the elements of GNSS as it pertains to civil aviation. The presentation addresses the following satellite navigation systems: the U.S. Global Positioning System (GPS), Russian GLONASS, European Galileo, Chinese Compass, Japanese Quasi Zenith Satellite System, and Indian Regional Navigation Satellite System. Nonlinear Statistical Modeling of Speech Speaker: Dr. Joseph Picone, Professor, Mississippi State University Abstract: In this presentation, we review our recent work on applying principles of nonlinear statistical modeling to acoustic modeling in speech recognition. Our goal is to improve recognition performance in noisy environments. We will discuss the use of an extended feature vector containing features based on correlation dimension, correlation entropy and Lyapunov exponents. We will also introduce a new acoustic model based on a probabilistic mixture of autoregressive models. Image and Video Quality Assessment: The Truth About PSNR Speaker: Dr. Amy Reibman, AT&T Bell Labs, IEEE Fellow Abstract: This talk provides a broad overview of objective methods for image and video quality assessment. We give visual examples and describe scenarios in which PSNR is misleading, inappropriate, or completely inapplicable. We also describe scenarios in which PSNR has proved very effective, where dramatic visual improvements in image quality can be achieved with its use. Finally, we present a sampling of alternate approaches to characterize image and video quality, including our recent contributions on measuring video quality inside the network. Waveform Diversity Techniques for Communications & Sensing Systems Speaker: Dr. Michael Picciolo, SAIC Technical Fellow Abstract: Communications and sensing systems transmit energy in the form of waveforms (e.g., wireless communications, radar, sonar, etc.). Local interference sources often corrupt the received waveform leading to bit errors or false detections or missed detections. Traditional waveforms have been designed for white noise environments and often perform poorly in colored noise (i.e. interference rich) environments. This talk will explain why the choice of waveform is fundamentally arbitrary and can therefore be optimized for and adapted to the signal environment in which it resides. We show examples for wireless communications and radar scenarios. We include cross-ambiguity function analysis to quantify performance tradeoffs. Automated and Adaptive Modeling, Detection, Prediction, and Control Speaker: Dr. Wallace E. Larimore, President, Adaptics Abstract: Over the past several decades, there has been a revolution in the modeling of linear Gaussian dynamic processes leading to more accurate and reliable detection, prediction, filtering and control. This presentation is focused on concepts at the foundation of this revolution - the use of reduced-rank statistical methods. Reduced Rank Adaptive Signal Processing Speaker: Dr. J. Scott Goldstein, IEEE Fellow Abstract: This talk introduces recent advances and algorithms for adaptive detection and estimation. In particular, emphasis is placed on how rank reduction can assist in successfully processing low powered signals in complicated colored noise signal environments. Examples will be presented from a multidisciplinary perspective. Radar Horizons Speaker: Dr. Joseph Guerci, IEEE Fellow Abstract: This talk provides a comprehensive survey of major new developments in radar research and development, from next generation intelligent signal processing to “super antennas” and radars that detect through buildings and around corners. The talk is designed to be of value to both the practicing radar engineer as well as non-specialists interested in advanced signal processing and systems engineering. Much of the material is drawn from Dr. Guerci’s own research—particularly from his recent 7 year term at the Defense Advanced Research Projects Agency (DARPA). Specific topics covered include: advanced STAP and knowledge-aided processing; waveform diversity and optimal MIMO radar; low-power density apertures for airship, space, and ground-based applications; and building penetration radar. Smart Camera Systems: A Technology Roadmap Speaker: Dr. Bruce Flinchbaugh, Texas Instruments (TI) Fellow Abstract: Consider a smart camera to be a software-programmable camera in which video data digitized from an image sensor is fully exposed to software for processing. In this talk we review the technology trends of programmable processors used in millions of smart cameras today. We consider the application-specific requirements of real-time image, video and vision processing in camera systems, emphasizing consumer electronics, automotive vision and video surveillance equipment. Finally, the requirements and trends are extrapolated to project future smart camera systems, as well as related challenges for vision research. Development of ZnO/SiO2/Si guided shear mode surface acoustic wave (SAW) devices for biosensor applications Speaker: Soumya Krishnamoorthy, PhD candidate in Electrical Engineering, University of Maryland College Park Abstract: Zinc Oxide (ZnO) is a material system with a highly reactive surface and offers the opportunity for effective bio-ZnO interfaces, thus making ZnO an excellent template for mass based bio-sensing applications. One of the critical steps in developing such devices is to functionalize specific proteins onto ZnO. In our work, we have immobilized a pro-inflammatory cytokine, namely, (Interleukin6) IL-6, in the range of 0.276 pg/ml-10 pg/ml, on the surface of ZnO and visualized at each stage with SEM and AFM studies. The protein-protein interactions were measured with the antigen/antibody immunoassay of solid-phase (Enzyme Linked Immunosorbent Assayt) ELISA. ZnO with a high piezoelectric coefficient is capable of generating very high frequency (GHz) surface acoustic wave devices. We have developed a ZnO/SiO2/ Si based high frequency guided shear mode surface acoustic wave device operating as high as 1.5 GHz. The mass sensitivities of the system have been modeled and experimentally verified. We find that the mass sensitivity that can be achieved in this system is more than double that seen in a Poly Methyl Meta Acrylate (PMMA) guiding layer based device. This SAW system has been used to detect Il-6 in trace amounts of a few fg of mass. Non-Conventional Image Formation Inspired By Opposing Neural Pathways Speaker: Dr. Damon Tull, co-founder and president, DVIP Multimedia, Inc. Abstract: In this talk we inspire the need to reform the image formation strategies of present day digital imaging systems. The current digital image formation strategies, inherited from film photography, have tradeoffs that allow image distortions to corrupt the final image and limit image utility after capture. Recent studies in biological image formation reveal mechanisms that predict and prevent image distortions. These mechanisms are expected to have a significant impact on many critical image processing tasks. DVIP Multimedia has begun to capture these mechanisms in a class of adaptive algorithms and, in this seminar, the impact of these algorithms in the area of image restoration is demonstrated. We will conclude with a discussion of future directions. Converting MATLAB Algorithms to FPGA or ASIC Designs Speaker: Dr Michael Bohm, CTO, Vice President, AccelChip Abstract: In the DSP domain, MATLAB is the domain-specific language of choice with 97% of DSP design implemented on dedicated DSP processors. MATLAB provides both an efficient system-level verification environment and an efficient path to implementation. Unfortunately, the process of converting MATLAB to "C" code to run on the processor is reaching its limits. A DSP processor's inherent limitation of serial operation is becoming a bottleneck for advanced high-performance algorithms. To solve this problem, a new methodology must be in place to convert algorithmic MATLAB to a register-transfer language (RTL) that can be used by industry-standard synthesis and verification tools. Companies that use the new methodology will benefit from greater productivity, both in terms of the domain-specific language and from the new breed of best-in-class tools they will enable. Signal Processing for Ocean Acoustic Tomography Speaker: Kathleen E. Wage, Assistant Professor, ECE Department, George Mason University Abstract: Ocean acoustic tomography uses measurements of the properties of acoustic signals to infer information about the ocean environment. Tomographic applications rely on a variety of signal processing techniques to acquire and analyze the data. For example, beamforming of sensor array data is used to separate signals that have taken different propagation paths through the environment. This talk provides an overview of tomographic signal processing, using illustrations from several long-range experiments in the North Pacific. Examples of sonar signals acquired during last summer's SPICEX deployment cruise will be presented. Current Techniques in Estimation Theory Speaker: Michael Tinston, SAIC Abstract: In this talk we will focus on some interesting results in estimation of Gaussian signals in Gaussian noise. First, the "optimal" solution will be derived for the minimum mean square error estimator. Next we will extend the results to the binary and multiple hypothesis estimation problems; this requires determination of the hypothesis concurrently with the calculation of the estimator. Finally, we will discuss the effect of realistic errors in the covariance of the signals. The techniques described will be highlighted with simple simulations.