EEMCS

Full text not available from this repository.

Abstract

The performance of modern wireless communication systems employing quadrature modulation is partly limited due to IQ imbalance. IQ imbalance essentially means that the two branches in the analog part of a quadrature receiver have different gain and phase characteristics causing finite image rejection.
Imperfections in technology and the realization in a design, caused by, for example, variations
in component values, ASIC process tolerances, and aging are the main causes for IQ imbalance. Due to these imperfections the quadrature mixers have gain and phase mismatches
between the in-phase and quadrature-phase branches of the quadrature receiver and the low-pass image-reject filters have different frequency characteristics.
This report is the result of a master thesis assignment for Electrical Engineering at the University of Twente for the chair of Signals and systems. Motivated by the results in [Valkama et al. (2001a)] the goal of this assignment is to acquire in-depth knowledge about the IQ imbalance problem in quadrature receivers and to investigate possible solutions
for solving the (frequency-selective) IQ imbalance problems.
The frequency-selective IQ imbalance can be cancelled in the digital domain by using digital
signal processing techniques. Based on [Valkama et al. (2001a)] the IQ imbalance is compensated using an adaptive interference canceller where the coefficients of the adaptive
filter are updated using an LMS algorithm. A general but thorough signal analysis is carried out on the IQ imbalance problem and the adaptive interference canceller is evaluated
and compared with the optimal Wiener filter solution.
It is shown that the adaptive interference canceller solution can be used to improve the image rejection of a quadrature receiver with IQ imbalance as described in [Valkama et al. (2001a)] by 60dB. This (impressive) result, using the traditional power-normalized LMS algorithm, is then further improved with respect to the convergence speed. A new algorithm, the correlation-based variable-step-size LMS (or CV-LMS for short), is developed
and it is shown that this new algorithm improves the convergence speed by 27%.

Export this item as:

• BibTeX
• XML

To correct this item please ask your editor

Repository Staff Only: edit this item