Proceedings of the KIEE Conference (대한전기학회:학술대회논문집)
- 2003.11c
- /
- Pages.685-688
- /
- 2003
A Consideration on ML Blind Signal Estimation based on Finite-Alphabet Characteristic in QPSK Modulation
QPSK 신호 입력시스템에서의 유한 알파벹 기반 ML 블라인드 신호 추정 비교
Abstract
In this paper, a performance comparison between two blind signal estimation algorithms in a LTI channel is considered. The two algorithms, Iterative Least-Squares with Projection (ILSP) and a modified ILSP, are based on the finite-alphabet property of input symbols. This case typically arises in a multiple access system with a sensor array antenna at the receiving end. We start with the formulation of a maximum-likelihood (ML) estimation problem under an additive white Gaussian noise assumption. A blind ML estimator is derived with its iterative algorithm for calculation. Then we narrow down the consideration of this problem to QPSK case so that a modified algorithm is proposed for