Espace membre

Cet espace est dédié aux étudiants, aux enseignants et au personnel administratif de l'école

Valider

Mot de passe oublié?

Actualités de l'école

31/03/2015

DOCTORATE THESIS DEFENSE of Ms. Leïla Nasraoui


Doctorate thesis defense on Saturday April 4th, 2015 at 9:00 a.m in Amphi at Iset’Com


Entitled : Advanced Synchronization Techniques for OFDM Systems

Presented by : Ms. Leïla Nasraoui 





Committee


Chair

Prof. Ridha Bouallegue

Professor at SUP’COM, Tunisia

 

 

 

Reviewers

Prof. Slimane Ben Slimane

Professor at KTH, Sweden

 

Prof.Noureddine Hamdi

Professor at INSAT, Tunisia

 

Examiner

Dr. Ines Kammoun Jemal

Associate professor at ENIS, Tunisia

 

Supervisor

Prof. Mohamed Siala

Professor at SUP’COM, Tunisia

 

Co-Supervisor

Dr. Leïla Najjar Atallah

Associate professor at SUP’COM, Tunisia


Abstract


The Orthogonal Frequency-Division Multiplexing (OFDM) is a largely adopted multi-carrier modulation technique in recent broadband wireless communication standards for its ability to cope with severe channels. OFDM systems are however very sensitive to synchronization errors that destroy the orthogonality between sub-carriers resulting in an important performance degradation. This thesis takes an overall look at reduced-complexity data-aided synchronization techniques for OFDM systems, by trading between performance and computational load in possibly fast varying wireless transmission environments.


Exploiting a preamble with specific repetitive structure, we propose a robust Brute-Force (BF) single-stage synchronization technique, based on differential correlation, which provides high detection accuracy at the expense of a huge computational load. To overcome this disadvantage, we propose the Reduced-Complexity (RC) two-stage technique that splits the synchronization processing into a coarse and a fine stages. The first stage, based on a sliding correlation, determines a short uncertainty interval over which the second fine stage, based on differential correlation, is carried. The combined use of the sliding correlation, characterized by its low complexity, and the differential correlation which is much more complex, carried over a short interval length around the coarse estimate, results in an overall reduced-complexity approach. To ensure the relevance of both proposed BF and RC approaches, we evaluate them in the standards IEEE 802.11a/g in which the signal structure allows to directly apply the proposed approaches.


For more optimized performance, we study the design of the preamble training sub-sequence defined respectively in the time and frequency domains. The first issue gives an important insight on how the binary training sub-sequence (length and type) should be chosen in order to achieve a better trade-off between detection accuracy and complexity. The second issue concerns a further complexity reduction through replacing the differential correlation operations by simple sign changes, using quantified training sub-sequences obtained as inverse Fourier transform of binary sequences defined in the frequency domain.


Furthermore, based on the transmission of a Zadoff-Chu (ZC) sequence as temporal preamble two new synchronization Simply-Differential (SD) and Doubly-Differential (DD) approaches are proposed. Both of the SD and DD approaches are proved to allow accurate detection of the ZC sequence start. Based on the fact that the LTE standard provides dedicated ZC sequences for the primary synchronization, the proposed approaches are successfully adapted and applied to primary synchronization in the LTE standard, which includes symbol start detection and sector identifier search.


Finally, we propose a new synchronization scheme for 2x1 MISO-OFDM systems, which exploits the full spatial diversity provided by differential Alamouti Space-Time Block Coding (STBC) for non-coherent synchronization and thus avoids any channel estimation requirement.


Keywords


OFDM, Data-aided synchronization, Time offset, Frequency offset, Binary PN sequences, Zadoff-Chu sequences, Differential correlation, MISO, Differential STBC.