Share Email Print

Proceedings Paper

Design and analysis of supervised and decision-directed estimators of the MMSE/LCMV filter in data-limited environments
Author(s): Jeffrey M. Farrell; Ioannis N. Psaromiligkos; Stella N. Batalama
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

In this paper we quantify theoretically the effect of the desired-signal power level on the mean square filter estimation error and the normalized output signal-to-interference-plus-noise-ratio (SINR) of sample matrix inversion (SMI)-type estimates of the minimum mean-square-error (MMSE) and the linearly constrained minimum variance (LCMV) filters. We prove that in finite data support situations filters that utilize a sample average estimate of the desired-signal-absent input correlation matrix exhibit superior normalized filter output SINR and mean square filter estimation error when compared to filters that utilize a sample average estimate of the desired-signal-present input correlation matrix. Finally, we investigate pilot-assisted and decision-directed adaptive filter implementations that exhibit near desired-signal-absent SMI-filtering performance while they are trained using desired-signal-present data/observations.

Paper Details

Date Published: 23 July 2003
PDF: 11 pages
Proc. SPIE 5100, Digital Wireless Communications V, (23 July 2003); doi: 10.1117/12.487947
Show Author Affiliations
Jeffrey M. Farrell, SUNY/Buffalo (United States)
Ioannis N. Psaromiligkos, McGill Univ. (Canada)
Stella N. Batalama, SUNY/Buffalo (United States)

Published in SPIE Proceedings Vol. 5100:
Digital Wireless Communications V
Raghuveer M. Rao; Soheil A. Dianat; Michael D. Zoltowski, Editor(s)

© SPIE. Terms of Use
Back to Top