header-about

EECS500 Spring 2017 Department Colloquium

Presenter: 
Arjuna Madanayake
Title: 
Analog Acceleration of Linear and Non-Linear Partial Differential Equation (PDE) Systems: Mathematical Development and Circuit Models (STTR Phase- I, DARPA)
Affiliation: 
University of Akron
Location: 
White 411
Time: 
11:30 AM - 12:30 PM
Date: 
April 11, 2017

Analog radio-frequency (RF) transistors having hundreds of GHz of bandwidth may offer an intriguing solution to computational physics problems that require extensive numerical simulation. Such scientific problems typically involve systems of linear and non-linear partial differential equations (PDEs). A favorite example is the solution of mathematical models used in simulating magnetohydrodynamics in fusion reactors, such as Tokamaks. In this DARPA project, we are interested in exploring analog computing, effectively returning to a pre-digital computation era (1930s) while using modern high-frequency analog devices to investigate whether it would be possible to obtain computational accelerations over conventional supercomputers. This talk reports on fundamental mathematics, algorithms and analog circuit architectures that have been developed in during phase-I (first year) of this research. Continuous-time solution of Maxwell’s Equations will be discussed in detail using electromagnetics as a linear example, followed by some ideas on how the analog computing approach can be extended to non-linear PDE systems, such as plasma dynamics, with emphasis on RF integrated circuit based implementation for highest possible speeds. The talk will explore concepts such as passive and active circuit synthesis, microwave and mm-wave active circuits and analog-digital hybrid computing for scientific model acceleration. Design examples using general immitance converter (GIC) based simulators will be explained. Finally, future collaborative exploration (phase-II) involving CWRU will be highlighted. Additional plans and opportunities for collaboration between UA and CWRU will be discussed.

Biography: 

Dr. Arjuna Madanayake completed the B.Sc. in Electronic and Telecommunication Engineering from the University of Moratuwa, Sri Lanka, with first class honors, and the MS and Ph.D. degrees, both in Electrical Engineering, from the University of Calgary, Canada. He joined the Department of Electrical and Computer Engineering at the University of Akron (UA) as a tenure-track Assistant Professor in 2010. He is now an Associate Professor, and leads the Advanced Signal Processing Circuits Group. His research interests are in multi-dimensional signal processing, digital and analog circuits and systems, wireless communications, antenna arrays, and FPGA/VLSI systems for real-time digital signal processing. His research is currently supported by two awards from DARPA, an award from Office of Naval Research, and three awards from the National Science Foundation. His is the university PI of an STTR Phase-II (due to start in Fall 2017) from the DARPA ACCESS Program (Defense Sciences Office) in collaboration with PI Dr. Dale Mugler, Ocius Technologies, LLC, Co-PI Dr. S.I. Hariharan (UA) and Co-PI Dr. Soumyajit Mandal (CWRU).