Dspace de universite Djillali Liabes de SBA >
Thèse de Doctorat en Sciences >
Electronique >

Please use this identifier to cite or link to this item: http://hdl.handle.net/123456789/2223

Titre: On Bit-Interleaved Coded Modulation with Iterative Decoding.
Auteur(s): BENMAHMOUD, Slimane
Encadreur: DJEBBARI, Ali
Mots-clés: BICM
MAP decoder
Iterative Decoding
Turbo codes
BCJR algorithm
Viterbi algorithm
Issue Date: 19-Apr-2018
Résumé: To increase the diversity order of coded modulation, Zehavi proposed Bit Interleaved Coded Modulation (BICM). Compared to Trellis Coded Modulation (TCM), BICM is very flexible in the design of the encoder and the modulator. BICM outperforms Ungerboeck’s TCM over fully interleaved Rayleigh fading channel but it suffers degradation over AWGN channel. BICM with soft-decision feedback significantly outperforms TCM and performs closely to Turbo-TCM over both AWGN and Rayleigh fading channel. This thesis considers BICM with iterative decoding (BICM-ID) in its soft-decision feedback version. In this work, we’ve studied the Soft-Input Soft-output (SISO) decoding algorithm and we’ve investigated the modulation block to improve the adaptability and the performance of this system: first, after a detailed theoretical study on BICM, its performance has been characterized. Then, to get a clear idea on iterative decoding and to fully explore the BCJR algorithm used later in BICM-ID, turbo codes have been studied. After this, the focus was on the joint optimization of signal constellation points and their mappings where a new improved symbol mapper/8-ary constellation was introduced. Finally, we’ve investigated the MAP algorithm’s variants. In particular, we’ve proposed a new approximation for the Jacobian logarithm and we’ve demonstrated its suitability for MAP decoding in BICM-ID.
Description: Doctorat en sciences
URI: http://hdl.handle.net/123456789/2223
Appears in Collections:Electronique

Files in This Item:

File Description SizeFormat
D_ELN_Slimane_Benmahmoud.pdf27,79 MBAdobe PDFView/Open
View Statistics

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.


Ce site utilise la plate-forme Dspace version 3.2-Copyright ©2014.