Mitch Haspel
Date: September 24, 2008 (Wednesday)
Time: 11:45am (refreshment starts at 11:30am)
Location: ECEC 202, NJIT

About the Presenter:

Mitch Haspel was granted the Ph.D. degree in electrophysics/electronics engineering for research into the interaction between CEM microwaves and CEF electron beams, and the masters and bachelors degrees in electronics engineering, all from the City University of New York. His global interests lie in the analysis and determination of design solutions for electronics related problems such as communications/telecommunications and signal processing systems utilizing mathematical and computer software oriented techniques and algorithms. He serves as Reviewer and Member of Technical Program Committee for IEEE ICACT conference in Korea during the past several years as well as session chair for mobile communications in 2005. He was selected as a referee and reviewer for the IEEE Journal on Selected Areas in Communications: Special Issue on Spread Spectrum for Global Communications, 1999. He has over 20 publications, some of the most recent in the fields of spread spectrum communications and CDMA, and has several patents, awarded and pending, and a technique cited in a recent US patent. He has worked continuously in R&D in a range of industrial institutions and has taught in a variety of academic environments. He is presently engaged in R&D and consulting at Stochastikos Solutions, R&D, which he founded. He is a member of Eta Kappa Nu and IEEE.

About the Talk:

This work extends the DOVETAIL technique for construction of families of complex spreading sequences for use in multiuser asynchronous spread spectrum, SS/CDMA, communication, to extraction of derived real-valued spreading sequence families, intended for similar application. It is shown here that the preferred characteristics of the complex family, in terms of probability of error, are partially carried over to the real and imaginary components when considered to be real-valued sequence, and intermediate improvement in probability of error is retained. Clearly the use of real-valued sequences results in reduced complexity in implementation. The improvement in probability of error is gauged with reference to two Gaussian approximations from the literature for the interference limited case, as well as with reference to control experiment randomly drawn sequences constructed by coin-tossing and drawing from the Gaussian distribution. The focus on the interference limited case where noise is suppressed allows one to observe the improvement yielded by the technique. The major parameters of the system are N, the sequence length, and q, the family size. Results from simulations are presented, demonstrating enhanced performance. The Gaussian approximations used for comparison are shown to underestimate the probability of error for the control experiment sequences while the generated real sequences have augmented performance with the complex sequences superior.

For more information contact Marlene Toeroek (marlene.toeroek@njit.edu)

Click here for the seminar archive

Note: All ECE MS thesis defense and Ph.D. dissertation (proposal) defense are counted towards ECE791.