Project Descriptions

Research Project: RF Power Transmission to and Telemetry From a Waterborne Sonar Array
Faculty Mentor: Rajeev Bansal
Project description: As the size of a sensor array increases, the complexity of the associated power-and signal-distribution cables become a major issue both in terms of affordability and reliability. If the necessary power to the sensors in the aperture can be supplied in a wireless fashion and if the data collected by the sensors can be transmitted similarly without hardwiring, the deployment of large arrays will be greatly facilitated. The proposed research, being carried out collaboratively with Electric Boat and Lockheed Martin, is a proof-of-concept study for the assessment of the wireless operation of a waterborne sonar array. The objective of the study is to develop and design a wireless power-and-signal-distribution system for use with a small number (at least four) of acoustic sensors. In the preferred embodiment of the system, power will be supplied in the form of RF energy by a single transmitting source located in a chamber behind the sensor array. It will be received and converted to DC in a module associated with each sensor, and data transmitted by the sensors will be picked up by a single receiving antenna for subsequent demultiplexing.

Research project: Single-Photon Transmission In Optical Networks
Faculty Mentor: Bing Wang
Project description: Our research effort focuses on experimental demonstration of multi-user quantum cryptography system. There are also extensive simulation efforts to aid in the design of quantum network topologies as well as understanding the optimal system parameters. Potential undergraduate REU projects include the simulation of single photon transmission in various optical network systems. For the simulations we use a combination of Virtual Photonics and MatLab. Virtual Photonics, a commercial optical network simulation software, will be used to simulate multi-user network environments. MatLab, a mathematical programming software, will be used to simulate physical events in greater details. In addition to the simulations, the undergraduate students may also work with the graduate students in the experimental component of the research project. This fall, one of the undergraduate students in our lab, John Kelm, will give an oral presentation on the results of his summer research on optical packet switching at the 2003 IEEE Lasers and Electro-Optics Society annual meeting, Tucson, Arizona, November 2003.

Research project: Optical Analog-to-Digital Converter
Faculty mentor: Eric Donkor
Project description: The goal of this project is to design and implement an optoelectronic circuit for high-speed sampling and quantization of 5GHz RF signals. Fast optoelectronic switching circuits, actuated by femtosecond laser pulses, designed with fast photodiodes, perform the RF sampling. Additionally optoelectronic/electronic circuits are designed to demultiplex the high-data rate sampled RF signals, to lower data rates so that the sampled signals can be directly quantized by electronic circuitry to digital gray-code signals. The research project will offer participants experience in key technology areas including optics, fiber-optics, optoelectronic, and high-speed electronic VLSI system design. In addition, participants will be exposed to state-of-the-art techniques and instrumentation such as working with femto-second laser systems, and bit-error-rate test system.

Research project: Biomedical Ultrasound and Optical Imaging Systems
Faculty mentor: Quing Zhu
Project description: Imaging systems are important tools for disease detection and diagnosis, such as breast cancer detection and cardiovascular plaque detection. This project will focus on design and testing of several coherent and diffused wave optical imaging systems and ultrasonic imaging systems. In the design, the student will participate in printed circuit board lay out and component distributions. In the testing, the student will use electronic and optical test equipment to evaluate signal to noise ratio of the electronic channels as well as noise and interference of the channels. The student will also participate in experimental evaluations of the diagnostic systems using phantom materials to simulate cancer cell clusters.

Research project: Properties and applications of resonant photonic-crystal films
Faculty mentor: Robert Magnusson
Project description: Resonating spatially periodic layers of dielectrics or semiconductors have interesting properties and numerous potential applications. At a certain light wavelength and incidence angle, the leaky mode excited in the grating layer will add constructively or destructively to the fields in the surrounding regions and a reflection or transmission peak will appear. This project will focus on design of optical filters and biosensors using computational tools available in the Nanophotonics Device Lab. Numerical results will be obtained with rigorous-coupled wave analysis (RCWA) and with the finite-difference time domain method (FDTD) to yield quantitative information on resonance efficiencies, relative field strengths, and spatial extents associated with the near fields, for a given set of device structural parameters. The student will participate in experimental research involving actual filter and sensor fabrication and testing.

Research project: A Physics Based Device Model For Transistors With Reduced Feature Size for Advanced Circuit Design
Faculty mentor: A. F. M. Anwar
Project description: With decreasing feature size the sophistication of the simulation tools used in industry to design high frequency/low voltage circuits are lagging behind the advancements in growth and fabrication technologies in terms of modeling inadequacies and runtime efficiency. Advanced devices will use thinner oxides (~10Å), have very short channel lengths (sub-100Å) and reduced film thickness (~100Å). Gate tunneling current is one of the dominant current carrying mechanism through thin oxides, becoming increasingly complicated to estimate in the presence of defects and impact ionization and are not incorporated in available simulators. This project will provide the circuit designers a simulation tool that is based upon the extracted BSIM_X parameters obtained from the underlying physics of the constituting device. The objectives of the proposed research are as follows: (a) Calculate the gate current using quantum mechanical calculations by taking into account the effect of elastic/inelastic scattering and any hole current due to impact ionization (b) Model transport and charge control of short channel devices and (c), and (d) model quantum transport to obtain device characteristics. Temperature and strain dependent intrinsic parameters in the BSIM_X format will be extracted by solving Schroedinger, Poisson’s, Laplace’s and time dependent Boltzman equations, for use in circuit simulation and in the calculation of noise and non-linear circuit performance parameters (e.g. IP3, IMD etc.) in RF and analog and mixed signal circuits.

Research project: Scalable Network Storage for Content Delivery
Faculty mentor: John A. Chandy
Project description: Content delivery networks offer a mechanism to provide efficient access across a WAN to various forms of distributed content such as images, video, static data, etc. A critical problem in the design of these networks is the intelligent routing to the closest and latest copy of a piece of content. This project will focus on architecture that uses distributed nodes in a clustered storage network with a specially designed file system to handle content resolution, data mapping, load balancing, and redundancy. In order to direct clients to the closest copy of the data, a mechanism is needed to route requests to the appropriate node holding the data. While there are methods to do load balancing within a tightly defined LAN using some form of TCP/IP forwarding, these techniques do not translate to the WAN. WAN-based techniques such as HTTP client redirection are not appropriate for file transfer protocols such as NFS, CIFS, or FTP. Therefore, for this research project, clustered storage architectures will be extended to use a TCP/IP connection transfer mechanism. A connection transfer protocol will be designed in order to transfer control from one TCP endpoint to another, thus allowing data to flow directly from the server to client without passing through a load-balancer/redirection node. The client will not be aware that the connection has been transferred, as the new host server will service all requests as if it was the original end point. The undergraduate student assigned to this project will design algorithms for the connection transfer and implement software to enable TCP connection transfer for storage networks.

Research project: Optimal Routing of a Vessel in Presence of Currents
Faculty mentor: Yaakov Bar-Shalom
Project description: This research will develop a finite element algorithm for optimization of the route of a vessel between two given points in the presence of currents (of magnitude comparable to the speed of the vessel) that vary both in space as well as in time. The optimization criterion will be minimum travel time. The path between the two points can include “exclusion zones” where the vessel cannot enter. The currents are given on a space-time grid and they can change significantly during the travel period. The speed of the vessel in the water as a function of its heading is also given on a grid. The optimization procedure to be used is dynamic programming, which, starting from the destination point with a given arrival time, will find the fastest path to it from the region of interest.

Research project: Nanoelectronics using 25 nm gate SiGe FETs programmable interconnects and Mosaic Architecture
Faculty mentor: Faquir Jain
Project description: This project aims to develop programmable interconnects by merging novel quantum dot based floating-trap nonvolatile memory cells with nanochannels field-effect transistors (FETs). Quantum dot nonvolatile memory cell structures are grown using layer-by-layer self-assembly. Here, the trapped charge density at the dot-cladding interface can be controlled to obtain the desired operating voltages for the nonvolatile memory. This project involves device and circuit modeling, epitaxial growth, and fabrication/processing using self-assembly techniques. The student will work in the fabrication laboratory as well as in the simulation laboratory. In addition to CADENCE and SILVACO design tools, there exist significant in-house software for quantum-dot modeling, which will be applied in the project as well.

Research project: Hybrid Diagnostic Modeling Techniques
Faculty mentor: Krishna R. Pattipati
Project description: The existing diagnostic modeling techniques can be classified into four groups: analytical (quantitative), qualitative, graph-based dependency and rule-based models. For example, automotive engineers have found quantitative simulation to be a vital tool in the development of advanced control systems. This project will investigate the advantages and disadvantages of these modeling techniques as applied to automotive systems. The project will develop hybrid model-based techniques that seamlessly employ graph-based dependency models and quantitative (analytical) models for intelligent diagnosis. The student will develop MATLAB-SimulinkÔ based analytical models of engine, transmission, anti-lock brake and suspension systems for proof-of-concept demonstrations.

Research project: Agent-based Vehicle Health Management Architectures
Faculty mentor: Krishna R. Pattipati
Project description: This project is investigating various three-tier architectures for on-board and off-board vehicle health management. Specifically, we are investigating hierarchical (Subsystem-resident agent, Functional area agent, Vehicle expert agent) and totally distributed agent-based architectures. In the latter architecture, each model-based agent monitors its own local subsystem, and communicates with its neighbors and possibly a web-based supervisor agent when unusual events occur. A broker agent will resolve conflicts among agents and act as an intermediary between local and supervisory agents. The student will develop Java-based software modules for testing the health management architectures.

Research project: Adaptive cross-layer designs for high-speed wireless access
Faculty mentor: Shengli Zhou
Project description: The increasing user demand for mobile applications such as wireless web browsing and wireless video streaming continues to drive the technology towards high-rate ubiquitous information access. The ``bottleneck'' for such applications is often the wireless link, due to the scarce bandwidth and power resources, and poor performance in the presence of multi-path fading and interference. This project will focus on high-speed wireless access with an advanced physical layer based on adaptive modulation and coding (AMC), and further improve the system performance by exploiting the interplay between the physical and higher protocol layers. The wireless link will be first emulated on a wired-line platform, and the students will test the impact of cross-layer optimization by measuring the delay and throughput of the data traffic, as well as by video streaming demonstration. With programmable short-range wireless transceivers, such as IEEE802.11b and ultra-wideband (UWB), the effect of adaptive cross-layer optimization will then be illustrated in real-time hardware setup.

Research Project: Scheduling and coordination of overhaul and repair services
Faculty mentor: Peter Luh
Project Description: Many industries rely on the proper functioning of their key assets, such as airlines on jet engines, electric utilities on generators, army on helicopters, automobiles for families, and semiconductor manufacturers on photolithography machines. A major obstacle on operation efficiency is asset unavailability due to forced outages and regular maintenance operations. The latter, although planned, are generally characterized by long and uncertain durations. Instead of embarking on more intensive preventative maintenance programs, the trend is to monitor and analyze asset conditions, and perform maintenance operations based on conditions to increase asset availability and reliability. This imposes major challenges on the networks of maintenance service providers, including overhaul shops, repair shops, part distributors, and spare part manufacturers, to have short and predictable turn-around-times while reducing inventory in systems traditionally characterized by massive uncertainties.

The undergraduate students will work on scheduling and coordination of overhaul and repair service networks under dynamic and uncertain environments to meet the challenges of condition-based maintenance on discrete simulation and agent-based implementation to test and validate the methods and for the distribute mobile multi-agent implementation. Our goal is to reap the full benefit of the new maintenance service paradigm by having short and predictable turn-around-times while reducing inventory levels under a dynamic and uncertain environment, making the best use of asset condition information and the information technology infrastructure.

Research Topic: Crystallographic Characterization of ZnSeTe Heteroepitaxial Layers
Faculty mentor: John Ayers
Project Description: Heteroepitaxial semiconductors are of interest in a number of applications including optoelectronics, high-frequency electronics, and high-temperature sensors. Heteroepitaxial ZnSeTe on GaAs is of interest for green light emitting diodes (LEDs) and lasers, which are difficult to implement in the GaInN or AlInGaP material systems. The student involved in this project would learn to use a high-resolution x-ray diffractometer and measure diffraction profiles for ZnSeTe/GaAs (001) epitaxial layers, grown in our laboratory by Metalorganic Vapor Phase Epitaxy (MOVPE). The student would also be involved in the interpretation of the results, calculating the relaxed lattice constant, composition, strain (in-plane and out-of-plane) and the threading dislocation density. This project will help the student develop a broad understanding of semiconductor materials, their crystallography, and characterization.

Research Topic: Mission-Specific Integrated Microsystems
Faculty mentor: Lei Wang
Project Description: Mission-specific integrated microsystems such as surveillance sensory systems are continually pushing the need for minimizing power dissipation, increasing throughput, integrating complex functionality, extending operational life, and improving reliability. At the same time, these systems are subject to considerable constraints on size, weight, and power dissipation, and need to operate in a wide range of temperature and extreme environmental conditions. The design of these systems needs to be optimized with respect to a number of issues such as multiple real-time data channels, increasing communication bandwidth, and sufficiently long mission duration. This project will focus on robust design of high-performance mission-specific integrated microsystems. Interested undergraduate students will be involved in different aspects of the project such as simulation of advanced signal detection, estimation and recognition algorithms, algorithm transform to low-power/high-speed VLSI architectures, embedded microarchitecture design, chip synthesis and verification, and post-silicon testing.

Research Laboratory: Physiological Acoustics Laboratory
Faculty mentor: Monty A. Escabí
Project Description: At a cocktail party our brains effortlessly attend to a single speaker despite high levels of superfluous background clutter. Humans and mammals easily filter important biological signal such as speech and vocalizations amongst high levels of noise; yet artificial speech recognition systems using computer technologies fail miserably at this seemingly simple computational task. In our laboratory, we study how the central nervous system processes sound information. Our goal is to understand how single brain cells (neurons) and networks of neurons consolidate sound information from the environment into a single perceptual and cognitive experience. Over many years of evolution, the brains of mammals have evolved efficient coding strategies for dealing with computational problems that are relevant behaviorally, and we are interested in identifying such computational strategies. We do this by recording electrical activity from single neurons or by optical imaging various brain regions and developing quantitative models that describe the encoding performed by single neurons and/or brain regions. Aside from helping us develop a general theory for the “neuronal code”, such an understanding could significantly benefit the design of optimal speech recognition systems and could significantly improve preprocessing strategies for assistive hearing devices (e.g., hearing aids and cochlear implants).

Research Topic: Dynamic Hardware/Software Co-management of Mobile Computers for Low Energy
Faculty mentor: Yunsi Fei
Project Description: Modern mobile systems have become an integral part of the lives of millions of users. The design of such systems imposes several new challenges, as it must consider demanding and dynamic resource requirements and stringent constraints. For mobile systems, energy becomes a first-class resource due to the weight/size limits for the battery. The objective of the proposed work is to investigate and develop a methodology and tool to improve the energy efficiency of mobile systems from both the hardware and software perspectives simultaneously.

It is well-known that many multimedia applications have the ability of trading off output quality for system resources, and their peak and average demands differ greatly. This provides us with a huge opportunity for energy optimization. However, three new questions are posed for this opportunity. Would users like to sacrifice a little bit quality for much longer battery lifetime under certain circumstances? Is it possible that energy consumption can be reduced greatly without affecting the quality? How should the system be managed automatically to explore such opportunity? To tackle these issues, we view it is necessary to design a middleware software sitting between the user and the computer system to act as an “intelligent manager“ of the mobile computer. The user determines if he/she would like certain mobile applications to scale for prolonging battery life. Such information is transferred to the underlying software tool. The software tool will determine at run-time how these application programs can gracefully degrade to a lower quality-of-service (QoS) level to conserve energy, meanwhile, how the hardware components can be scaled to just meet the requirements from different applications or different application QoS levels without any quality penalty. The project will focus on developing and implementing the first dynamic hardware/software co-management framework. Interested undergraduates will be involved in various aspects of the project, including constructing middleware software, developing adaptation policy and algorithms, building scaling circuitry, and integrating them on a commercial mobile computer platform.