Performance degradation of telecommunication transceivers due to parameter estimation errors

Abstract

In this thesis we consider the performance degradation of three telecommunication transceivers with respect to parameters uncertainties. The major contribution of our analysis is the derivation of simple and informative approximations to the excess mean square error, which uncover the basic factors that determine the performance of each transceiver under the pres- ence of parameter uncertainties. In Chapter 1, we brie°y present the channel models and the transceiver structures used in this thesis. In Chapters 2 and 3, we consider the case of the Multiple-Input Multiple-Output (MIMO) °at fading channel and two widely known transceiver schemes, i.e., the trans- mit MIMO Wiener ?ltering and the Tomlinson-Harashima (TH) precoding. For the case of the transmit MIMO Wiener ?ltering, degradation is due to channel estimation errors and time-variations, and noise second-order statistics estimation errors. For the case of the TH precoder, degradation is due to channel estimation errors and time-variations. For both cases, our ?nal expressions uncover the factors that determine the performance degradation in practice, and the relative importance between di®erent error sources. In Chapter 4, we consider the case of the Single-Input Single-Output (SISO) frequency selective channel and the minimum mean square error (MMSE) linear equalizer. The error sources that cause the system degradation is the channel and CFO estimation errors. Our aim is to uncover the relative importance of these error sources. It turns out that the CFO estimation error is much more important than the channel estimation error. This fact leads to useful conclusions concerning the optimal training sequence design for joint CFO and channel estimation.