Title

Finite Arithmetic Effect On The I/Q Channel Mismatch Correction Of A Direct Conversion Wireless Receiver Employing Fast Least Mean Square Algorithm

Keywords

Adaptive and fixed-point numbers; Direct Conversion Receiver; Fixed Point Fast Least Mean Square; I/Q Mismatch

Abstract

The field of wireless communications has grown tremendously over the last decade and wireless communication receivers have been a major field of research. Superheterodyne Receivers are one of the most widely used receivers and they require passive components thus not making a ideal solution for small portable receivers needed in cell phones and wireless LAN applications. Direct Conversion Receiver (DCR) is a feasible solution for small, low power devices but they have I/Q mismatch associated with it. The I/Q mismatch severely degrades the performance of DCR. Adaptive solutions using Fast Least-Mean-Square algorithm (FLMS) has been proposed using floating-point arithmetic and successfully implemented to over come this mismatch. In this paper, a fixed-point approach for implementing this algorithm has been proposed because of the advantages offered like reduced hardware cost, size and complexity. In the Quadrature Phase Shift Keying (QPSK) receivers, the imbalance caused due to the mismatch in low pass filter (LPF) and local oscillator (LO) are considered for the simulation. Fixed Point Fast Least Mean Square (FPFLMS) algorithm is implemented successfully to over come this mismatch. This method has a near optimum solution if the proper word lengths are used for the fixed-point implementation. The word lengths have to be mapped taking in consideration the vector that is to be mapped. Mapping individual vectors according to the range and using a good precision obtain results comparable to that obtained by using a floating-point implementation.

Publication Date

12-1-2005

Publication Title

WMSCI 2005 - The 9th World Multi-Conference on Systemics, Cybernetics and Informatics, Proceedings

Volume

5

Number of Pages

175-180

Document Type

Article; Proceedings Paper

Personal Identifier

scopus

Socpus ID

84867370013 (Scopus)

Source API URL

https://api.elsevier.com/content/abstract/scopus_id/84867370013

This document is currently not available here.

Share

COinS