Fast channel estimation using maximum-length shift-register sequences

Document Type

Conference Proceeding

Publication Date

1-1-2005

Abstract

Channel state information (CSI) is crucial at the MAC layer for wireless mobile networks. Therefore, fast channel estimation is in demand for high scalability communication networks. Least-square channel estimation techniques are widely adopted in the majority communication standards. However, they usually involve the large-dimensional matrix inversion, whose heavy computational complexity cannot be extendable for long channel filters. Since the maximum-length shift-register (MLSR) sequences, or m-sequences, possess the well controlled second-order cyclic statistics, they have been used as the training sequences for least-square channel estimators. In this paper, we analyze the statistical characteristics of m-sequences and design a corresponding computationally efficient channel estimation algorithm. We also evaluate the performance of the proposed algorithm in terms of mean square error and compare the computational complexity between our proposed algorithm and the conventional scheme. It can be justified that our proposed algorithm can achieve both efficiency and satisfactory performance for communication channel estimation.

Publication Source (Journal or Book title)

IEEE Vehicular Technology Conference

First Page

1897

Last Page

1900

This document is currently not available here.

Share

COinS