You are on page 1of 9

Seminar report: EECS Department

FALL-2014

DEPENDABILITY METRICS
By,
Dr.Mansoor Alam,
Chair of EECS Department
University of Toledo
August 25th, 2014
Dr. Alam initiated his talk with the goals and requirements for students to pursue their
Masters program and then continued with the topic called Dependability Metrics.
Fig: Dependability Metrics
The dependability metrics such as availability, reliability, safety, confidentiality, integrity,
maintainability were introduced. The simple defining forms for these dependability metrics were
introduced as:
The readiness for usage leads to availability;
The continuity of service leads to reliability;
The non-occurrence of catastrophic consequences on the environment leads to safety;
The non-occurrence of unauthorized disclosure of information leads to confidentiality;
The non-occurrence of improper alterations of information leads to integrity;
The aptitude to undergo repairs and evolutions leads to maintainability.
The talk was then followed by addressing few terms as error, fault, failure which are inter
related. Finally the seminar concluded by explaining the thesis outline and research tools.

TOWARDS A NEW GENERATION OF PROGRAMMING LANGUAGES


THE UNIVERSITY OF TOLEDO
1

Seminar report: EECS Department


FALL-2014

By,
Dr. Henry F. Ledgard
Professor
Department of Electrical Engineering & Computer Science
The University of Toledo
September 8th, 2014
Dr. Henry F. Ledgard has received his B.A. from Tufts University in 1964 and his Ph.D.
from the Massachusetts Institute of Technology in 1969. He has been at the University of Oxford
as a post-doctoral fellow. He was a faculty member at Johns Hopkins University, and
subsequently was on the faculty at the University of Massachusetts/Amherst. In 1977, he became
a member of the design team to create the new programming language ADA. From 1979 he
conducted his own consulting and writing practice. In 1989 he joined the faculty at the
University of Toledo. Presently he is dealing with Software Specification and Design course. He
wrote books like Programming language landscape, professional software volume I and II etc., in
coordination with others.
Dr. Henry Ledgard started his presentation with an Idea to explore an exciting new
principle for programming language design. This principle is known as the Separation
Principle. It is derived from a language and language environment called VisiSoft produced by
Prediction Systems in New Jersey. The use of the separation principle allows one to use a
graphical approach to modular design, reduces program complexity and considerably reduces
scope-rules. He made use of an English-like syntax as well as the idea named rules.
Dr. Ledgard stated that both concepts are also inspired by VisiSoft. These ideas suggest a
new direction for the programming language field. Dr. Ledgard also discussed the limitations of
conventional languages, the difficulty with Object- Oriented Programming, and a fundamental
problem in the field that many software ideas are based on thoughtful opinions that are not
necessarily supported by data. He concluded by explaining the key role played by programming
language development these days.

THE UNIVERSITY OF TOLEDO


2

Seminar report: EECS Department


FALL-2014

Fig: Design and Architecture of VISISOFT Programming Language

THE UNIVERSITY OF TOLEDO


3

Seminar report: EECS Department


FALL-2014

DEVELOPMENT OF PARALLEL ARCHITECTURES


FOR SIGNAL PROCESSING APPLICATIONS
By,
Dr. Mohsin M. Jamali
Professor
Department of Electrical Engineering & Computer Science
The University of Toledo
September 15th, 2014
Dr. Mohsin M. Jamali does research in the area of parallel computer architectures for
real time applications. He has developed parallel architectures for on-board processors, signal,
and radar and sensor array processing applications. Dr. Jamalis research has been sponsored by
NASA, Office of Naval Research, Department of Energy and Air Force. He has worked as
Summer Research Faculty at the Wright Patterson Air Force Base in Dayton, Ohio. He
collaborates with researchers from Bowling Green State University and Air Force Research
Laboratory at WPAFB, Dayton, Ohio.
This talk was presented on the development of parallel processing systems which can be
designed using the Multi-core processors, Graphics processing units, Multi-core DSPS and field
programmable arrays. Initially he started with Moores Law, detailed description of Multi-core,
its advantages and challenges were explained. Then the talk continued with the exploration of
parallelism which is explained with the help of IBM cell BE processor, play station 3 diagrams.
The talk is carried forward with the explanation of Graphic processing unit flow, wide band
receiver control system, certain algorithms, the simulation and results. The presentation gave an
idea of what is Field programmable gate array and how a CAD Tool for Parameterized FPGA
Based FFT architecture can be used, with the help of CAD Tool Software Flow graph.

Fig: Java GUI communicating with both VHDL library and Xilinx (software tool)

THE UNIVERSITY OF TOLEDO


4

Seminar report: EECS Department


FALL-2014

The second part of this talk was focused on the software tool that is capable of
auto-generating a fully optimized VHDL which allows the designer to focus on the
overall Soc performance. This can be used for FPGA/ASIC design. Later he spoke about
the implementation of target tracking, location estimation on GPUs and software defined
radar test bed followed by the multidimensional research problem such as radar
monitoring, IR monitoring, acoustics etc. The research will help in identifying behavior
of nocturnally active birds and others. This talk was concluded after demonstrating the
practical implementation of the application

THE UNIVERSITY OF TOLEDO


5

Seminar report: EECS Department


FALL-2014

EM-ANNs
FOR
COMPUTER
AIDED
MODELLING
OPTIMIZATION: FROM THEORY TO PRACTICE

AND

By
Dr. Vijaya Kumar Devabhaktuni
Professor
Department of Electrical Engineering & Computer Science
The University of Toledo
September 22th, 2014
Dr.Vijay Devabhaktuni received the B.Eng. degree in EEE and the M.Sc. degree in
Physics both from BITS, India, in 1996, and the Ph.D. in Electronics from Carleton University,
Canada, in 2003. He held the 2005-2008 Canada Research Chair in Computer-Aided HighFrequency Modeling and Design at Concordia University, Montreal, Canada. He is currently Full
Professor in the EECS Department at the University of Toledo. His R&D interests include
applied electromagnetics, biomedical applications, computer aided design, device modeling,
infrastructure monitoring, neural networks, optimization, RF/microwave devices, and virtual
reality.
Computer aided design (CAD) lies at the heart of electronics industry. For instance, wireless
markets that rely on CAD tools reached $22 billion in 2003 and are growing at 15% per year.
Recent drive in the microelectronics and RF/microwave industries toward GHz frequencies
demands efficient and powerful computer aided modeling, simulation, and optimization tools.
Circuits operating at GHz frequencies emulate full-wave electromagnetic (EM) behaviors.
Although traditional CAD tools allow virtual design of high-frequency circuits (e.g. RFICs), the
resulting designs may not work, because high-frequency libraries of such tools typically lack EM
flavors. The seminar covered educational, fundamental concepts of artificial neural networks
(ANNs) based high-frequency modeling techniques striving for judicious use of CPU-expensive
EM data. From an industry perspective, approaches to smart integration of high-frequency
models into CAD-tool libraries will be briefly described. Several examples in EMC/EMI and
RF/microwave domains will also be covered. Should the time permit, the talk will highlight the
speakers recent transition into interdisciplinary arenas including biomedical engineering,
enhancing the access to radio spectrum, structural health monitoring (SHM), and national
defense.

THE UNIVERSITY OF TOLEDO


6

Seminar report: EECS Department


FALL-2014

APPLIED OPTIMIZATION TECHNIQUES


TRACKING and DATA TRANSMISSION

FOR

DIM

OBJECT

By
Dr.Ezzatollah Salari
Professor
Department of Electrical Engineering and Computer Science
September 29th, 2014
Dr. Ezzatollah Salari received his B.S. from Iran University of Science and Technology
in Tehran, M.S. and Ph. D. degrees in Electrical Engineering from Wayne State University. He is
involved in teaching and research in the areas of image/video/signal processing, pattern
recognition & neural networks, and data compression. He has published Over 100 papers (about
40 journal papers) and has taught over 30 new courses in both the Electrical Engineering &
Computer Science areas. His research papers are very often cited in the literature (e.g., the work
on motion estimation in video was cited 564 times at this point). He has supervised 9 Ph.D.
dissertations and more than 25 M.S. theses. He served as Graduate Director of the EECS
Department at UT from 2000 to 2005. Dr.Salari also worked as a research fellow at NASA
Langley, Goddard and Lewis (Glenn) research centers during the summers of 1987, 1988, 1990
and 1991 on various projects. He worked on funded projects sponsored by NASA, Edison
Center, and more recently on projects sponsored by MIOH-UTC, U.S. Department of
transportation.
The Talk begun with the introduction to the Optimization Technique for the dim object
tracking & data transmission. This technique is a key in the area of the image and video analysis.
The various applications such as surveillance, search and rescue, remote sensing, floating mine
detection, and traffic monitoring also use the above technique and technology.
An effective and fast algorithm to detect and track low observable targets in a digital image
sequence was presented in this seminar.
The method is based on the Genetic Algorithm (GA) along with a multi-stage hypothesis
testing scheme. In addition, for the purposes of the robust transmission of image data over
packet-switched networks, a GA based technique which provides an optimized data packet
scheme for the transmission of SPIHT coded bit streams will be presented. If time permits, an
optimized method for obtaining super-resolution images using an integrated neural networks will
also be discussed.

THE UNIVERSITY OF TOLEDO


7

Seminar report: EECS Department


FALL-2014

POSITION SENSOR LESS CONTROL OF PERMANENT


MACHINES AT LOW SPEED

MAGNET

By
Dr. Arshan Khan
Lead Motor Calibration Engineer
Fiat Chrysler Automobiles
Detroit, Michigan
October 6th, 2014
Dr. Arshan Khan is currently working as lead motor calibration engineer at Fiat
Chrysler Automobiles. He received his PhD in Electrical Engineering (specialization in motor
controls) from Florida International University (FIU), Miami in 2009 and his M.S. in Electrical
Engineering (specialization in Power and Controls) from Illinois Institute of Technology in 2006.
He worked as postdoctoral fellow at Energy Systems research lab, FIU Miami for before joining
automotive industry. He worked at Power and Control design INC., LA, California for a year
before joining Chrysler in 2011 as a motor control engineer. He is IEEE Senior Member and his
research interests include power electronics control of electric machines, electric/hybrid vehicle
drives, wind energy, sensor less drives and smart grid. Dr. Khan holds one US patent (another
patent submitted and pending approval) and has published over 15 peer-reviewed technical
papers.
The main theme about this discussion was about the sensor less control of permanent
magnet machine at low speed. The permanent magnets are gaining it importance and popularity
in the industry these days due to high torque to volume, high power factor and low maintenance
cost. These are primary requirement for many applications. In order to control the permanent
magnet motor, rotor position determination is essential.The position sensor However increases
the cost of drive system and decreases the reliability, which is unacceptable for most
applications. This reason why the usage of sensorless technology for Permanent Magnets drive
applications is becoming more and more popular.
In this seminar, different sensorless control methods were discuused for different speed
regions.Estimators were proposed to obtain rotor position information. Several assistive
algorithms, including an online observer parameter adaption scheme, a model reference adaptive
system based speed estimator, and an estimated speed-based oscillation mitigation scheme, were
proposed to improve the performance of the rotor position estimation and the sensorless
Permanent Magnets control system. The proposed methods were effective for both salient-pole
and nonsalient-pole Permanent Magnets. Permanent-magnets are widely used in industrial
THE UNIVERSITY OF TOLEDO
8

Seminar report: EECS Department


FALL-2014

applications owing to their distinctive advantages, such as high efficiency, high power density,
and wide constant power region.

THE UNIVERSITY OF TOLEDO


9