Turbo égalisation de faible complexité avec estimation des canaux multi trajets à évanouissements rapides
|Abstract:||Beyond their capacity limitations due to the transmission channels efects, nowadays, some embarked multimedia applications such as video telephony, broadband internet, etc., require ecient digital processing of the received signals. Turbo detection is a promising method to address these requirements. It allows exploiting all the available information at the receiver. However, its performance for the severe channels is highly dependent on the transmission parameter estimation. Its computational complexity is mainly due to the SISO modules. In this context, this thesis study the design and analysis of low complexity iterative receiver for the coded symbols over unknown fast fading multipath Rayleigh channels. In the first part of the thesis, we investigate the SISO turbo equalization for single user : we propose and analyze an iterative low complexity architecture, including equalization, decoding and channel estimation with its statistics. Using semi analytical tools, we explain the propagation efect on the turbo equalizer and justify the choice of modules realizing the best complexity/performance compromise. Furthermore, we demonstrate that the proposed architecture is efective in severe channel conditions where the impulse response and statistics are unknown to the receiver. In the second part, our research focuses on multi-user turbo detection for uplink asynchronous DS-CDMA systems. Two situations are considered. For the first one, the receiver has a single antenna. We propose and evaluate an iterative architecture involving lower complexity modules, handling multi-user detection, decoding and channels estimation. As for the second situation, we use the receiver diversity principle and we extend the proposed architecture modules to the multi antennas context. Simulation results showed that the proposed architecture eliminate almost all the interference caused by the channel without increasing significantly the required signal to noise ratio.|
|Document Type:||Thèse de doctorat|
|Open Access Date:||17 April 2018|
|Collection:||Thèses et mémoires|
All documents in CorpusUL are protected by Copyright Act of Canada.