No description available.

Le codage réseau est un paradigme qui permet une utilisation efficace du réseau. Il maximise le débit dans un réseau multi-saut en multicast et réduit le retard. Dans cette thèse, nous concentrons notre attention sur l’intégration du codage réseau aux applications multimédias, et en particulier aux systèmes avancès qui fournissent un service vidéo amélioré pour les utilisateurs. Nos contributions concernent plusieurs scénarios : un cadre de fonctions efficace pour la transmission de flux en directe qui utilise à la fois le codage réseau et le codage par description multiple, une nouvelle stratégie de transmission pour les réseaux sans fil avec perte qui garantit un compromis entre la résilience vis-à-vis des perte et la reduction du retard sur la base d’une optimisation débit-distorsion de l'ordonnancement des images vidéo, que nous avons également étendu au cas du streaming multi-vue interactive, un système replication sociale distribuée qui, en utilisant le réseau codage en relation et la connaissance des préférences des utilisateurs en termes de vue, est en mesure de sélectionner un schéma de réplication capable de fournir une vidéo de haute qualité en accédant seulement aux autres membres du groupe social, sans encourir le coût d’accès associé à une connexion à un serveur central et sans échanger des larges tables de métadonnées pour tenir trace des éléments répliqués, et, finalement, une étude sur l’utilisation de techniques de séparation aveugle de source -pour réduire l’overhead encouru par les schémas de codage réseau- basé sur des techniques de détection d’erreur telles que le codage de parité et la génération de message digest. / Network coding is a paradigm that allows an efficient use of the capacity of communication networks. It maximizes the throughput in a multi-hop multicast communication and reduces the delay. In this thesis, we focus our attention to the integration of the network coding framework to multimedia applications, and in particular to advanced systems that provide enhanced video services to the users. Our contributions concern several instances of advanced multimedia communications: an efficient framework for transmission of a live stream making joint use of network coding and multiple description coding; a novel transmission strategy for lossy wireless networks that guarantees a trade-off between loss resilience and short delay based on a rate-distortion optimized scheduling of the video frames, that we also extended to the case of interactive multi-view streaming; a distributed social caching system that, using network coding in conjunction with the knowledge of the users' preferences in terms of views, is able to select a replication scheme such that to provide a high video quality by accessing only other members of the social group without incurring the access cost associated with a connection to a central server and without exchanging large tables of metadata to keep track of the replicated parts; and, finally, a study on using blind source separation techniques to reduce the overhead incurred by network coding schemes based on error-detecting techniques such as parity coding and message digest generation. All our contributions are aimed at using network coding to enhance the quality of video transmission in terms of distortion and delay perceived

Στην παρούσα διπλωματική εργασία μελετώνται μέθοδοι διαχωρισμού σημάτων σε ανεξάρτητες συνιστώσες. Αφού δοθεί ο ορισμός του προβλήματος και μια αναφορά στις κυριότερες μεθόδους για την αντιμετώπισή του, γίνεται σαφές ότι δεν μπορούν να σχεδιαστούν γενικές μέθοδοι διαχωρισμού σημάτων. Παρά την πληθώρα των πρακτικών προβλημάτων στα οποία βρίσκει εφαρμογή το μαθηματικό πρότυπο, δεν είναι δυνατός ο σχεδιασμός μιας ενιαίας μεθόδου που να αντιμετωπίζει αποτελεσματικά όλες τις περιπτώσεις διαχωρισμού σημάτων. Ο αναγνώστης πληροφορείται για τις περιοχές έρευνας και ανάπτυξης των διαφόρων μεθόδων καθώς και για τις εφαρμογές τους σε διάφορους τομείς της σύγχρονης επιστήμης. Στη συνέχεια, υλοποιούνται κάποιες από αυτές τις μεθόδους και παρουσιάζονται τα αποτελέσματα προσομοίωσης πραγματικών πειραματικών δεδομένων που λήφθηκαν για την εκπόνηση της συγκεκριμένης διπλωματικής εργασίας. Τα αποτελέσματα εξάγονται με την χρήση και υλοποίηση αλγόριθμου επεξεργασίας των δεδομένων στο πρόγραμμα Matlab και μελετώνται εκτενέστερα με το πρόγραμμα Adobe Audition 1.5. Τέλος, παρουσιάζονται τα συμπεράσματα από την εφαρμογή του αλγόριθμου στα πραγματικά δεδομένα και δίνεται μια μαθηματική- θεωρητική βάση για την βελτιστοποίηση των μεθόδων διαχωρισμού σημάτων. / -

Separating unknown signal mixtures into their constituent parts is a difficult problem in signal processing called blind source separation. One of the benchmark problems in this area is the extraction of the fetal heartbeat from an electrocardiogram in which it is overshadowed by a strong maternal heartbeat. This thesis presents a study of a signal separation technique called independent component analysis (ICA), in order to assess its suitability for the maternal-fetal ECG separation problem. This includes an analysis of ICA on deterministic, stochastic, simulated and recorded ECG signals. The experiments presented in this thesis demonstrate that ICA is effective on linear mixtures of known simulated or recorded ECGs. The performance of ICA was measured using visual comparison, heart rate extraction, and energy, information theoretic, and fractal-based measures. ICA extraction of clinically recorded maternal-fetal ECGs mixtures, in which the source signals were unknown, were successful at recovering the fetal heart rate.

This report intends to verify the possibility that the FastICA algorithm could be applied to the GPS system to eliminate the impulsive noise from the receiver end. As the impulsive noise is so unpredictable in its pattern and of great energy level to swallow the signal we need, traditional signal selection methods exhibit no much use in dealing with this problem. Blind Source Separation seems to be a good way to solve this, but most of the other BSS algorithms beside FastICA showed more or less degrees of dependency on the pattern of the noise. In this thesis, the basic mathematic modelling of this advanced algorithm, along with the principles of the commonly used fast independent component analysis (fastICA) based on fixed-point algorithm are discussed. To verify that this method is useful under industrial use environment to remove the impulsive noises from digital BPSK modulated signals, an observation signal mixed with additive impulsive noise is generated and separated by fastICA method. And in the last part of the thesis, the fastICA algorithm is applied to the GPS receiver modeled in the SoftGNSS project and verified to be effective in industrial applications. The results have been analyzed. / 6

This dissertation introduces three novel multiuser detection approaches in Multiple Input Multiple Output (MIMO) Orthogonal Frequency Division Multiplexing (OFDM) systems by blind signal separation (BSS) techniques. The conventional methodologies for multiuser detection have to retransmit channel state information (CSI) constantly from the transmitter in MIMO ODFM systems at the cost of economic efficiency, because they require more channel resources to improve the communication quality. Compared with the traditional methodologies, the proposed BSS methods are relatively efficient approaches without the unnecessary retransmission of channel state information. The current methodologies apply the space-time coding or the spatial multiplexing to implement an MIMO OFDM system, which requires relatively complex antenna design and allocation in the transmitter. The proposed Spatial Division Multiple Access (SDMA) method enables different mobile users to share the same bandwidth simultaneously in different geographical locations, and this scheme requires only one antenna for each mobile user. Therefore, it greatly simplifies the antenna design and allocation. The goal of this dissertation is to design and implement three blind multiuser detection schemes without knowing the channel state information or the channel transfer function in the SDMA-based uplink MIMO OFDM system. The proposed scenarios include: (a) the BSS-only scheme, (b) the BSS-Minimum Mean Square Error (MMSE) scheme, and (c) the BSS-Minimum Bit Error Ratio (MBER) scheme. The major contributions of the dissertation include: (a) the three proposed schemes save the commercially expensive cost of channel resources; (b) the proposed SDMA-based uplink MIMO OFDM system simplifies the requirements of antennas for mobile users; (c) the three proposed schemes obtain high parallel computing efficiency through paralleled subcarriers; (d) the proposed BSS-MBER scheme gains the best BER performance; (e) the proposed BSS-MMSE method yields the best computational efficiency; and (f) the proposed BSS-only scenario balances the BER performance and computational complexity.

<p>A Real-Time Brain-Computer Interface is a technical system classifying increased or decreased brain activity in Real-Time between different body movements, actions performed by a person. Focus in this thesis will be on testing algorithms and settings, finding the initial time interval and how increased activity in the brain can be distinguished and satisfyingly classified. The objective is letting the system give an output somewhere within 250ms of a thought of an action, which will be faster than a persons reaction time. </p><p>Algorithms in the preprocessing were Blind Signal Separation and the Fast Fourier Transform. With different frequency and time interval settings the algorithms were tested on an offline Electroencephalographic data file based on the "Ten Twenty" Electrode Application System, classified using an Artificial Neural Network. </p><p>A satisfying time interval could be found between 125-250ms, but more research is needed to investigate that specific interval. A reduction in frequency resulted in a lack of samples in the sample window preventing the algorithms from working properly. A high frequency is therefore proposed to help keeping the sample window small in the time domain. Blind Signal Separation together with the Fast Fourier Transform had problems finding appropriate correlation using the Ten-Twenty Electrode Application System. Electrodes should be placed more selectively at the parietal lobe, in case of requiring motor responses.</p>

A Real-Time Brain-Computer Interface is a technical system classifying increased or decreased brain activity in Real-Time between different body movements, actions performed by a person. Focus in this thesis will be on testing algorithms and settings, finding the initial time interval and how increased activity in the brain can be distinguished and satisfyingly classified. The objective is letting the system give an output somewhere within 250ms of a thought of an action, which will be faster than a persons reaction time. Algorithms in the preprocessing were Blind Signal Separation and the Fast Fourier Transform. With different frequency and time interval settings the algorithms were tested on an offline Electroencephalographic data file based on the "Ten Twenty" Electrode Application System, classified using an Artificial Neural Network. A satisfying time interval could be found between 125-250ms, but more research is needed to investigate that specific interval. A reduction in frequency resulted in a lack of samples in the sample window preventing the algorithms from working properly. A high frequency is therefore proposed to help keeping the sample window small in the time domain. Blind Signal Separation together with the Fast Fourier Transform had problems finding appropriate correlation using the Ten-Twenty Electrode Application System. Electrodes should be placed more selectively at the parietal lobe, in case of requiring motor responses.