Tewks Web Site

Current Courses

Vita

Facilities Development

Research

Photography

ECE Department

School of Engineering & Science

Stevens Institute of Technology
Laboratory for Advanced Information and Computational Systems (LAICS)
(S. Tewksbury, Director)
The Laboratory for Advanced Information and Computation Systems (LAICS) was established as a laboratory facility within the Microelectronic Systems Research Center. The LAICS focuses on system-level research, including the underlying principles of system design and optimization and the demonstration of such principles through prototype realizations. The LAICS was founded by Prof. Tewksbury, who served as its director until 1998. Prof. Hornak was associate director. Additional information is provided in a separate LAICS report.

LAICS established a library of information resources (a link is included here to illustrate the approach taken to provide students with supporting information) for graduate and undergraduate students involved in LAICS-related research.

The laboratory facilities are shown in the figure below.

RESEARCH THRUSTS OF LAICS

Research thrusts are generally in the area of integration of high performance systems (including low-cost personal products as well as high end professional products) in which the co-integration of advanced, state-of-the-art computing components (silicon CMOS VLSI integrated circuits), state-of-the-art programmable VLSI components (FPGAs, CPLDs, etc.) for customization of a full system using commercial parts, software design and insertion for contemporary programmable VLSI components, and software development environments for such systems. Primary thrust areas are as follows.
High Performance Computational Accelerators

This program addresses the prototyping of advanced computational accelerators, drawing on state-of-the-art, commercial digital signal processors (DSPs) and innovative communications techniques for interconnecting several DSPs and associated memory and other support ICs into a versatile computational engine. Customization of the system is based on FPGA and other programmable VLSI components.

The "Intelligent Network of DSPs" project funded by the NSF illustrates one of the topics pursued to identify advanced architectures for multi-stage image processing. The general approach is described in
  • Communications-enhanced DSP Integrated Circuits such as the Analog Devices SHARC DSP and the TI 'C40 , 'C80, and 'C6x DSP families. Most of the research has drawn upon the SHARC DSP.

  • Cluster-based arrays, with overall architectures organized as parallel sets of clusters and with each cluster a parallel organization of DSPs ICs, memory, network nodes, and other supporting components. Typically, there are 4 to 8 DSPs per cluster. The SHARC DSP was particularly effective in allowing construction of DSP arrays without requiring significant additional "glue logic."

  • Software development environments exploiting data-flow representations to partition software elements among the available DSP clusters and, within a cluster, among the DSPs of the cluster.

  • Reconfigurable communications interfaces among the DSPs of a cluster.

  • Advanced interconnection technologies and protocols for alignment of communications capabilities with the needs of algorithms being executed.

  • Impact of advanced packaging (including passive MCM-D and active substrate MCM-D technologies) for realization of high performance clusters.
Advanced Image Processing Systems and Environments
Advanced image processing systems and environments exploiting high performance parallel computational accelerators are the focus of this area. A primary emphasis was in the general area of medical imaging applications.

This program highlights the following representative topics.
  • Low level image processing algorithms and their efficient realization on cluster-based DSP arrays. A focus is on communications aspects of image file movement and distribution to the multiple processors.

  • Non-local image processing algorithms for image enhancement and analysis. A focus is on the impact of non-local algorithms on the efficient realization of local algorithms driven by non-local parameters.

  • Feature extraction within parallel computational accelerators, with the features extending across processor/memory boundaries within a parallel computational accelerator.

  • Higher level image processing algorithms including image overlay, image presentation, etc.

  • Image-flow based communications as a foundation for establishing "image file" access, manipulation, and replacement within a distributed/shared memory organization.
Information Systems
The evolving activities in this topic highlight distance learning and medical informatics.
  • Distance Learning: The CEMR Laboratory for Computer-Based Instructional Technology (LCIT) is closely supported by the LAICS. Representative examples of topics for distance learning include the following.
    • Frameworks for Web-based educational modules including interactive, asynchronous "discussions" with the teacher.

    • New capabilities for image presentation including approaches for local generation of complex graphical information based on the minimum required information from the server.

    • Integration of Virtual Laboratories within Web-based course modules, allowing experimentation related to the topic of the module.

  • Medical Informatics: This area, under development, highlights two coupled topics, namely medical informatics and medical imaging & image processing environments. Representative activities include the following:
    • NSF/EPSCoR Research Cluster: Medical Imaging and Image Processing. This cluster initiative includes researchers from the Robert C. Byrd Health Sciences Center, the Dept. of Electrical and Computer Engineering at WVU, and the College of Engineering at WVUIT. Initial funding will acquire a parallel computer from Silicon Graphics for collaborative and individual research projects related to the broad area of medical image systems, presently highlighting PET, CT, and related imaging systems.

    • Collaborative Medical Informatics Program. This initiative, funded by equipment grants from Hewlett Packard, established a large database system which models the evolving medical informatics system of health science and medical organizations, providing external researchers with a source of medical information (images, data records, etc.) through which research by a wide range of researchers can be pursued. Facilities for research related to home health care topics were also established with this program This is a collaborative initiative among the Robert C. Byrd Health Sciences Center, the Concurrent Engineering Research Center, the Dept. of Electrical and Computer Engineering of WVU, and the Dept. of Statistics and Computer Science of WVU