Channel order estimation is an important problem in many fields including signal processing, communications, acoustics, and more. In this thesis, blind channel order estimation problem is considered for single-input, multi-output (SIMO) FIR systems. The problem is to estimate the effective channel order for the SIMO system given only the output samples corrupted by noise. Two new methods for channel order estimation are presented. These methods have several useful features compared to the currently known techniques. They are guaranteed to find the true channel order for noise free case and they perform significantly better for noisy observations. These algorithms show a consistent performance when the number of observations, channels and channel order are changed. The proposed algorithms are integrated with the least squares smoothing (LSS) algorithm for blind identification of the channel coefficients. LSS algorithm is selected since it is a deterministic algorithm and has some additional features suitable for order estimation. The proposed algorithms are compared with a variety of dierent algorithms including linear prediction (LP) based methods. LP approaches are known to be robust to channel order overestimation. In this thesis, it is shown that significant gain can be obtained compared to LP based approaches when the proposed techniques are used. The proposed algorithms are also compared with the oversampled single-input, single-output (SISO) system with a generic decision feedback equalizer, and better mean-square error performance is observed for the blind setting. Channel order estimation problem is also investigated for semi-blind systems where a pilot signal is used which is known at the receiver. In this case, two new methods are proposed which exploit the pilot signal in dierent ways. When both unknown and pilot symbols are used, a better estimation performance can be achieved compared to the proposed blind methods. The semi-blind approach is especially effective in terms of bit error rate (BER) evaluation thanks to the use of pilot symbols in better estimation of channel coecients. This approach is also more robust to ill-conditioned channels. The constraints for these approaches, such as synchronization, and the decrease in throughput still make the blind approaches a good alternative for channel order estimation. True and effective channel order estimation topics are discussed in detail and several simulations are done in order to show the significant performance gain achieved by the proposed methods.

Κανάλια με αραιή κρουστική απόκριση εμφανίζονται πάρα πολύ συχνά σε εφαρμογές ασύρματων κυρίως τηλεπικοινωνιακών συστημάτων. Παραδείγματα τέτοιων εφαρμογών είναι η εκπομπή HDTV (HighDefinitionΤelevision) ή εκπομπή μέσω υποθαλλάσιων ακουστικών καναλιών. Σε όλες αυτές τις εφαρμογές η μορφή του καναλιού διαμορφώνεται από το φαινόμενο της πολυδιόδευσης. Συνεπώς ο δέκτης λαμβάνει έναν περιορισμένο αριθμό από διαφορετικές εκδοχές του εκπεμπόμενου σήματος καθεμία με διαφορετική εξασθένιση και καθυστέρηση. Ως εκ τούτου η συνάρτηση της κρουστικής απόκρισης ενός τέτοιου καναλιού αποτελείται από ελάχιστα μη μηδενικά στοιχεία σε συγκριση με το μήκος της, καθένα από τα οποία αντιστοιχεί σε ένα από τα μονοπάτια πολυδιόδευσης. Για την ισοστάθμιση αυτών των καναλιών έχουν προταθεί διάφορες τεχνικές, πολλές από τις οποίες εκμεταλλεύονται την ιδιαίτερη αυτή μορφή της κρουστικής απόκρισης. Πολλοί από τους προτεινόμενους ισοσταθμιστές καναλιών απαιτούν την παρεμβολή ακολουθίων εκμάθησης ανάμεσα στην ακολουθία δεδομένων, οι οποίες είναι εκ των προτέρων γνωστές στον δέκτη. Χρησιμοποιούνται δε προκειμένου ο αλγόριθμος εκτίμησης του καναλιού να συγκλίνει όσο το δυνατόν ταχύτερα στην επιθυμητή τιμή. Μειονέκτημα αυτών των μεθόδων είναι η επιβάρυνση του ωφέλιμου εύρους ζώνης που συνεπάγεται. Ωστόσο η εκ των προτέρων γνώση της αραιής μορφής της κρουστικής απόκρισης εχει δώσει αφορμή για την σχεδίαση ισοσταθμιστών με περιορισμένο μήκος αλλά εξίσου καλή απόδοση. Οι συμβατικές τεχνικές εκτίμησης καναλιών, όπως η Least Square μέθοδος, δεν εκμεταλλεύονται αυτή την γνώση. Οι πρόσφατες δε εξελίξεις στην ανακατασκευή αραιών σημάτων μέσω τεχνικών συμπιεσμένης καταγραφής (compressed sensing) έχουν οδηγήσει στην μελέτη της εφαρμογής τέτοιων τεχνικών στο πρόβλημα της εκτίμησης καναλιού. Η μέθοδος της συμπιεσμένης καταγραφής στηρίζεται στη δυνατότητα ανακατασκευής αραιών σημάτων από πλήθος δειγμάτων αισθητά κατώτερο από αυτό που προβλέπει το θεωρητικό όριο του Nyquist. Έχει αποδειχθεί ότι η ανακατασκευή αυτή είναι δυνατή όταν το σήμα ή έστω κάποιος μετασχηματισμός του περιέχει λίγα μη μηδενικά στοιχεία σε σχέση με το μήκος του. Οι εφαρμογές αυτών των τεχνικών εκτείνονται και σε άλλα πεδία όπως η επεξεργασία εικόνας, η μαγνητική τομογραφία, η ανάλυση γεωφυσικών δεδομένων, η επεξεργασία εικόνας radar, η αστρονομία κ.α. Στα πλαίσια αυτής της εργασίας παρουσιάζονταιοι βασικές αρχές που διέπουν την ανακατασκευή αραιών σημάτων μέσω της επίλυσης υποορισμένων συστημάτων γραμμικών εξισώσεων. Παράλληλα παρουσιάζονται οι κυριότεροι αλγόριθμοι που έχουν προταθεί για την υλοποίηση της και εξετάζονται ως προς την απόδοση και την υπολογιστική πολυπλοκότητα τους. Εν συνεχεία εξετάζεται η εφαρμογή αυτών των αλγορίθμων στο πρόβλημα της εκτίμησης αραιών καναλιών. Προτείνονται δε ισοσταθμιστές αραιών καναλιών βασισμένοι σε εκτιμητές απόκρισης που χρησιμοποιούν τεχνικές συμπιεσμένης καταγραφής. / Channels with sparse impulse response are very common in wireless telecommunications systems applications. Example of such channel is HDTV channel where multipath distribution of the transmitted signal results in a sparse form of the channel impulse response. Several different versions of the same signal are received, each one with its own gain and delay. As a result, channel impulse response has a few non zero taps compared to its length, its one corresponding to a different distribution path. Several techniques for estimating and equalizing such channels have been proposed, most of them taking advantage of this sparse form of the impulse response. The transmission of a training sequence known to the receiver is required for this purpose. It is used so that the channel estimation algorithm at the receiver converges faster. The disadvantage of the use of a training sequence is the fact that the useful bandwidth is reduced. However the a priori knowledge of the sparse form of the training sequence has led to the design of equalizers that require short training sequences but have satisfactory performance. Channel estimation techniques based on least square method do not take advantage of this idea. On the other hand recent progress on sparse signal reconstruction using compressed sensing techniques has led scientists to research the potential use of such algorithms in channel estimation. Compressed sensing is based on the idea of reconstructing a sparse signal using less samples that those predicted by Nyquist theorem. It has been proved that such a reconstruction is feasible if the reconstructed signal is sparse enough. In this dissertation several sparse signal reconstruction algorithms are presented and their performance and complexity are evaluated. Then the application of these algorithms on channel estimation equalization problem is analyzed.

Wireless underwater acoustic (UWA) communications is a developing field with various applications. The underwater acoustic communication channel is very special and its behavior is environment-dependent. The UWA channel is characterized by low available bandwidth, and severe motion-introduced Doppler effect compared to wireless radio communication. Recent literature suggests that machine learning (ML)-based channel estimation and equalization offer benefits over traditional techniques (a decision feedback equalizer), in UWA communications. ML can be advantageous due to the difficultly in designing algorithms for UWA communication, as finding general channel models have proven to be difficult. This study aims to explore if ML-based channel estimation and equalization as a part of a sophisticated physical layer structure can offer improved performance. In the study, supervised ML using a deep neural network and a recurrent neural network will be utilized to improve the bit error rate. A channel simulator with environment-specific input is used to study a wide range of channels. The simulations are utilized to study in which environments ML should be tested. It is shown that in highly time-varying channels, ML outperforms traditional techniques if trained with prior information of the channel. However, utilizing ML without prior information of the channel yielded no improvement of the performance.

Fixed wireless communication is a cost-efficient solution for flexible and rapid front-/backhaul deployments. Technologies including dual polarization, carrier aggregation, and higher order modulation schemes have been developed for enhancing its throughput. In order to better support the massive traffic increment during network evolution, novel wireless backhaul solutions with possible new dimensions in increasing the spectral efficiency are needed. Line-of-Sight (LoS) Multiple-Input-Multiple-Output (MIMO) communication is such a promising candidate allowing the throughput to scale linearly with the deployed antenna pairs. Spatial multiplexing with sub-channels having approximately equal quality exists within a single LoS direction. In addition, operating at millimeter wave (mmWave) frequencies or higher, the abundantly available bandwidth can further enhance the throughput of LoS MIMO communication. The mmWave LoS MIMO communication in this work exploits the spatial multiplexing from the structured phase couplings of a single path direction, while most of the state-of-the-art works in mmWave communication focus on the spatial multiplexing from the spatial signature of multiple path directions. Challenges: The performance of a LoS MIMO system is highly dependent on the antenna topology. Topologies resulting in theoretically orthogonal channels are considered as optimal arrangements. The general topology solution from a unified viewpoint is unknown. The known optimal arrangements in the literature are rather independently derived and contain restrictions on their array planes. Moreover, operating at mmWave frequencies with wideband signals introduces additional challenges. On one hand, high pathloss is one limiting factor of the received signal power. On the other hand, high symbol rates and relatively high antenna numbers create challenges in signal processing, especially the required complexity for compensating hardware imperfections and applying beamforming. Targets: In this thesis, we focus on antenna topologies and signal processing schemes to effectively handle the complexity challenge in LoS MIMO communications. Considering the antenna topology, we target a general solution of optimal arrangements on any arbitrarily curved surface. Moreover, we study the antenna topologies with which the system gains more streams and better received signals. Considering the signal processing, we look for low complexity schemes that can effectively compensate the hardware impairments and can cope with a large number of antennas. Main Contributions: The following models and algorithms are developed for understanding mmWave LoS spatial multiplexing and turning it into practice. First, after analyzing the relation between the phase couplings and the antenna positions in three dimensional space, we derive a channel factorization model for LoS MIMO communication. Based on this, we provide a general topology solution from a projection point of view and show that the resulting spatial multiplexing is robust against moderate displacement errors. In addition, we propose a multi-subarray LoS MIMO system for jointly harvesting the spatial multiplexing and array gains. Then, we propose a novel algorithm for LoS MIMO channel equalization, which is carried out in the reverse order w.r.t. the channel factorization model. The number of multiplications in both digital and analog implementations of the proposed solution is found to increase approximately linearly w.r.t. the number of antennas. The proposed algorithm thus potentially reduces complexity for equalizing the channel during the system expansion with more streams. After this, we focus on algorithms that can effectively estimate and compensate the hardware impairments. A systolic/pipelined processing architecture is proposed in this work to achieve a balance between computational complexity and performance. The proposed architecture is a viable approach that scales well with the number of MIMO streams. With the recorded data from a hardware-in-the-loop demonstrator, it is shown that the proposed algorithms can provide reliable signal estimates at a relatively low complexity level. Finally, a channel model is derived for mmWave systems with multiple widely spaced subarrays and multiple paths. The spatial multiplexing gain from the spatial signature of multiple path directions and the spatial multiplexing gain from the structured phase couplings of a single path direction are found simultaneously at two different levels of the antenna arrangements. Attempting to exploit them jointly, we propose to use an advanced hybrid analog/digital beamforming architecture to efficiently process the signals at reasonable costs and complexity. The proposed system can overcome the low rank property caused by the limited number of propagation paths.

This dissertation presents methods for modeling and mitigating voltage noise and timing jitter across high-speed chip-to-chip interconnects. Channel equalization and associated tuning schemes have been developed to target the distinct characteristics and signal degradation exhibited in the clock and data signals of multi-Gigabit/second digital communication links. Multiple methods for generating realistically degraded signals for the purpose of simulation are also presented and used to verify the proposed equalization and filtering topologies. Specifically, a new technique for modeling high-speed jittery clocks in the frequency domain is presented and shown to reduce transient simulation time and memory requirements, while simultaneously improving the timing resolution and accuracy of the simulation by minimizing the dependence on the transient simulation time-step. The technique is further developed to provide unprecedented control over the timing characteristics of the generated signals, and is then extended to the generation of random data signals with definable jitter statistics. Through these techniques,realistic clock and data waveforms are constructible, providing for the visualization of the combined effects of voltage and timing degradation, while at the same time tracking the phase relationship between the clock and data signals as they pass across their respective channels and through the receiving circuitry of the communication link. New methods for the automated tuning of second-order continuous-time channel equalizers are proposed based on the simulated or measured single pulse and double pulse responses of the transmission channel. Using only one degree of freedom, the methods target the reduction of inter-symbol interference (ISI) as identified in the single and double pulses. Through tuning either the circuit quality factor (Q), the peaking frequency, or the frequency zero, the methods are shown to adapt to a variety of channel lengths and datarates from the same original equalizer transfer function, implying a good degree of generality, while offering a simple, yet effective, method for ISI reduction. Finally, the design of an active 5 Gigahertz (GHz) bandpass filter, employed for high-speed clock conditioning, is presented and shown to address both random and deterministic components of the clock signal degradation. The bandpass transfer function is achieved through a combination of AC coupling and a resonant LC tank consisting of on-chip interleaved spiral inductors and a tunable capacitor array. Through adjusting the load capacitance in parallel with the inductors, the center frequency of the filter is tunable over a range of nearly 5GHz. The design targets a supply voltage of 1.2 volts and draws approximately 5.7 milliamps of current.

Bridging machine learning technologies to multiple-input-multiple-output (MIMO) communications systems is a primary driving force for next-generation wireless systems. This dissertation introduces a variety of neural network structures for symbol detection/equalization tasks in MIMO systems configured with two different waveforms, orthogonal frequency-division multiplexing (OFDM) and orthogonal time frequency and space (OTFS). The former one is the major air interface in current cellular systems. The latter one is developed to handle high mobility. For the sake of real-time processing, the introduced neural network structures are incorporated with inductive biases of wireless communications signals and operate in an online training manner. The utilized inductive priors include the shifting invariant property of quadrature amplitude modulation, the time-frequency relation inherent in OFDM signals, the multi-mode feature of massive antennas, and the delay-Doppler representation of doubly selective channel. In addition, the neural network structures are rooted in reservoir computing - an efficient neural network computational framework with decent generalization performance for limited training datasets. Therefore, the resulting neural network structures can learn beyond observation and offer decent transmission reliability in the low signal-to-noise ratio (SNR) regime. This dissertation includes comprehensive simulation results to justify the effectiveness of the introduced NN architectures compared with conventional model-based approaches and alternative neural network structures. / Doctor of Philosophy / An important topic for next-generation wireless systems is the integration of machine learning technologies with conventional communications systems. This dissertation introduces several neural network architectures to solve the transmission problems in wireless communications systems. The discussion focuses on the following major modern communications technologies: multiple-input-multiple-output (MIMO), orthogonal frequency-division multiplexing (OFDM), and orthogonal time frequency space (OTFS). In today's cellular networks, MIMO and OFDM are the major air-interface. OTFS is a novel technique that has been designed to work in a high-mobility setting. The implemented neural network structures are integrated with inductive biases of wireless communications signals and operate in an online training mode with limited training datasets. The neural network architectures, in particular, are based on reservoir computing, which is an efficient neural network computational system. A learning algorithm's inductive bias (also known as learning bias) is a collection of assumptions that the learner makes to infer outputs from unknown inputs. The dissertation introduces four different inductive priors from four different perspectives of MIMO communications systems. As a result, the neural network architectures can learn beyond observation and provide good generalization output in scenarios having model mismatch issues. The dissertation provides extensive simulation results to support the efficacy of the implemented NN architectures compared to alternative neural network models and traditional model-based approaches.

De nos jours, l'utilisation des drones ne cesse d'augmenter et de nombreuses études sont réalisées afin de mettre en place des systèmes de communication dronique destinés à des applications non seulement militaires mais aussi civiles. Pour le moment, les règles d'intégration des drones commerciaux dans l’espace aérien doivent encore être définies et le principal enjeu occupation est d'assurer une communication fiable et sécurisée. Cette thèse s’inscrit dans ce contexte de communication. Motivée par la croissance rapide du nombre des drones et par les nouvelles générations des drones commandés par satellite, la thèse vise à étudier les différents liens possibles qui relient le drone aux autres composants du système de communication. Trois principaux liens sont à mettre en place : le lien de contrôle, le lien de retour et le lien de mission. En raison de la rareté des ressources fréquentielles déjà allouées pour les futurs systèmes droniques, l'efficacité spectrale devient un paramètre crucial pour leur déploiement à grande échelle. Afin de mettre en place un système de communication par drones spectralement efficace, une bonne compréhension des canaux de transmission pour chacune des trois liaisons est indispensable, ainsi qu’un choix judicieux de la forme d’onde. Cette thèse commence par étudier les canaux de propagation pour chaque liaison : canaux de type muti-trajets avec ligne de vue directe, dans un contexte d’utilisation de drones à moyenne altitude et longue endurance (drones MALE). L’objectif de cette thèse est de proposer de nouveaux algorithmes de réception permettant d’estimer et égaliser ces canaux de propagation muti-trajets. Les méthodes proposées dépendent du choix de la forme d’onde. Du fait de la présence d’un lien satellite, les formes d’onde considérées sont de type mono-porteuse (avec un faible facteur de crête) : SC et EW-SCOFDM. L’égalisation est réalisée dans le domaine temporel (SC) ou fréquentiel (EW-SC-OFDM). L'architecture UAV prévoit l'implantation de deux antennes placées aux ailes. Ces deux antennes peuvent être utilisées pour augmenter le gain de diversité (gain de matrice de canal). Afin de réduire la complexité de l'égalisation des canaux, la forme d'onde EW-SC-OFDM est proposée et étudiée dans un contexte muti-antennes, dans le but d'améliorer l'endurance de l'UAV et d'accroître l'efficacité spectrale, une nouvelle technique de modulation est considérée: Modulation spatiale ( SM). Dans SM, les antennes de transmission sont activées en alternance. L'utilisation de la forme d'onde EW-SC-OFDM combinée à la technique SM nous permet de proposer de nouvelles structures modifiées qui exploitent l’étalement spectrale pour mieux protéger des bits de sélection des antennes émettrices et ainsi améliorer les performances du système. / Nowadays, several studies are launched for the design of reliable and safe communications systems that introduce Unmanned Aerial Vehicle (UAV), this paves the way for UAV communication systems to play an important role in a lot of applications for non-segregated military and civil airspaces. Until today, rules for integrating commercial UAVs in airspace still need to be defined, the design of secure, highly reliable and cost effective communications systems still a challenging task. This thesis is part of this communication context. Motivated by the rapid growth of UAV quantities and by the new generations of UAVs controlled by satellite, the thesis aims to study the various possible UAV links which connect UAV/aircraft to other communication system components (satellite, terrestrial networks, etc.). Three main links are considered: the Forward link, the Return link and the Mission link. Due to spectrum scarcity and higher concentration in aircraft density, spectral efficiency becomes a crucial parameter for largescale deployment of UAVs. In order to set up a spectrally efficient UAV communication system, a good understanding of transmission channel for each link is indispensable, as well as a judicious choice of the waveform. This thesis begins to study propagation channels for each link: a mutipath channels through radio Line-of-Sight (LOS) links, in a context of using Meduim Altitude Long drones Endurance (MALE) UAVs. The objective of this thesis is to maximize the solutions and the algorithms used for signal reception such as channel estimation and channel equalization. These algorithms will be used to estimate and to equalize the existing muti-path propagation channels. Furthermore, the proposed methods depend on the choosen waveform. Because of the presence of satellite link, in this thesis, we consider two low-papr linear waveforms: classical Single-Carrier (SC) waveform and Extented Weighted Single-Carrier Orthogonal Frequency-Division Multiplexing (EW-SC-OFDM) waveform. channel estimation and channel equalization are performed in the time-domain (SC) or in the frequency-domain (EW-SC-OFDM). UAV architecture envisages the implantation of two antennas placed at wings. These two antennas can be used to increase diversity gain (channel matrix gain). In order to reduce channel equalization complexity, the EWSC- OFDM waveform is proposed and studied in a muti-antennas context, also for the purpose of enhancing UAV endurance and also increasing spectral efficiency, a new modulation technique is considered: Spatial Modulation (SM). In SM, transmit antennas are activated in an alternating manner. The use of EW-SC-OFDM waveform combined to SM technique allows us to propose new modified structures which exploit exces bandwidth to improve antenna bit protection and thus enhancing system performances.

Terrestrial Digital Video Broadcasting (DVB-T) is a standard for wireless broadcast of MPEG-2 video. DVB-T is based on channel coding algorithms and uses Orthogonal Frequency Division Multiplexing (OFDM) as a modulation scheme. In this thesis, we have implemented the standard of ETSI EN 300 744 for Digital Video Broadcasting in MATLAB. This system is composed of the certain blocks which include OFDM modulation, channel estimation, channel equalization, frame synchronization, error-protection coding, to name a few of such blocks. We have investigated the performance of the complete system for different wireless broadcast impairments. In this performance analysis, we have considered Rayleigh fading multi-path channels with Doppler shift and framing synchronization errors and obtained the bit error rate (BER), and channel minimum square error performances versus different maximum Doppler shift values, different channel equalization techniques and different channel estimation algorithms. Furthermore, we have investigated different interpolations methods for the interpolation of channel response. It is shown that minimum mean-square error (MMSE) type equalization has a better performance in symbol estimation compared to zero forcing (ZF) equalizer. Also linear interpolation in time and low pass frequency interpolation, for time frequency interpolation of channel response can be used for practical application.

Submitted by Patrícia Cerveira (pcerveira1@gmail.com) on 2017-06-06T18:58:56Z No. of bitstreams: 1 Felipe Frasson- Dissertação.pdf: 984658 bytes, checksum: 05ae4f112679292aefe890dc2f563010 (MD5) / Rejected by Biblioteca da Escola de Engenharia (bee@ndc.uff.br), reason: Patrícia, o formulário de submissão apresenta vários erros, informações duplicadas e fora da formatação (orientador, coorientador, resumo, dentre outros). Atenciosamente, Catarina Ribeiro Bibliotecária BEE - Ramal 5992 on 2017-06-29T16:53:14Z (GMT) / Submitted by Patrícia Cerveira (pcerveira1@gmail.com) on 2017-06-29T19:32:38Z No. of bitstreams: 1 Felipe Frasson- Dissertação.pdf: 984658 bytes, checksum: 05ae4f112679292aefe890dc2f563010 (MD5) / Approved for entry into archive by Biblioteca da Escola de Engenharia (bee@ndc.uff.br) on 2017-07-03T13:00:12Z (GMT) No. of bitstreams: 1 Felipe Frasson- Dissertação.pdf: 984658 bytes, checksum: 05ae4f112679292aefe890dc2f563010 (MD5) / Made available in DSpace on 2017-07-03T13:00:12Z (GMT). No. of bitstreams: 1 Felipe Frasson- Dissertação.pdf: 984658 bytes, checksum: 05ae4f112679292aefe890dc2f563010 (MD5) / Em um sistema de comunica c~oes, os sinais s~ao transmitidos atrav es de canais de comunica c~ao que, idealmente, deveriam transportar os dados de maneira a n~ao causar distor c~ao alguma. Por em, em sistemas reais, existem limita c~oes que interferem neste processo causando degrada c~ao nas informa c~oes transmitidas, podendo comprometer sua recep c~ao. Tais limita c~oes ocorrem devido a presen ca de ru do aditivo, e principalmente por interfer^encia intersimb olica, esta caracterizada pela sobreposi c~ao de s mbolos gerados por uma mesma fonte transmissora. A equaliza c~ao de canal e uma das t ecnicas existentes que reduzem os efeitos da interfer^encia intersimb olica, dando maior con abilidade e robustez aos sistemas de comunica c~oes. Dentre as t ecnicas utilizadas para equaliza c~ao de canal, o uso de algoritmo adaptativos vem sendo amplamente utilizados devido as suas propriedades de se auto-ajustarem as varia c~oes que ocorrem ao longo do tempo. Este trabalho tem como objetivo veri car o comportamento de diferentes tipos de algoritmos adaptativos cegos ou semicegos, assim denominados por n~ao utilizarem sequ^encias de treinamento, aplicados a equaliza c~ao de canais esparsos. Canais esparsos s~ao encontrados em diversos sistemas de comunica c~oes como, por exemplo, na comunica c~ao sem o (telefonia m ovel, transmiss~ao de r adio e TV), ou, ainda, em canais subaqu aticos. Os algoritmos foram escolhidos com base em recentes estudos desta aplica c~ao, que operam em modo cego ou semicego e utilizam estat sticas de alta ordem, como os algoritmos Bussgang e Matching Pursuit. Os algoritmos foram implementados em ambiente de simula c~ao computacional no qual foram utilizados canais esparsos simples e de resposta ao impulso conhecida, permitindo comparar o comportamento dos diferentes algoritmos, em termos do sinal recuperado, e da inversa da resposta ao impulso do canal original. / In communications systems, information signals are transmitted through communications channels that, ideally, are delivered without distortions. However, on real communications channels there are limitations that interferes on the process, reducing the probability to recover the original signal at receiver. These distortions are basically thermal noise and Intersymbol Interference (ISI), caused by superposition on the received symbols received from the same source. Channel Equalization acts reducing these distortions, bringing more reliability to communications systems. The objective of this work is to verify di erent adaptive algorithms behavior, applied to sparse channel equalization problem. Many communications systems have sparse channels, like broadcast radio, television, mobile telephony and underwater communications. The selected algorithms used in this work includes high order statistics algorithms family, like Bussgang and Matching Pursuit. This kind of algorithms are widely used, with high relevance, for blind channel equalization. The selected algorithms were submitted to computer simulations using simple sparse channels and knowledge about their impulse response, in order to analyze their behavior in therms of bit error rate and the inverse impulse response of the channel.

Online learning algorithms have several key advantages compared to their batch learning algorithm counterparts: they are generally more memory efficient, and computationally mor efficient; they are simpler to implement; and they are able to adapt to changes where the learning model is time varying. Online algorithms because of their simplicity are very appealing to practitioners. his thesis investigates several online learning algorithms and their application. The thesis has an underlying theme of the idea of combining several simple algorithms to give better performance. In this thesis we investigate: combining weights, combining hypothesis, and (sort of) hierarchical combining.¶ Firstly, we propose a new online variant of the Bayes point machine (BPM), called the online Bayes point machine (OBPM). We study the theoretical and empirical performance of the OBPm algorithm. We show that the empirical performance of the OBPM algorithm is comparable with other large margin classifier methods such as the approximately large margin algorithm (ALMA) and methods which maximise the margin explicitly, like the support vector machine (SVM). The OBPM algorithm when used with a parallel architecture offers potential computational savings compared to ALMA. We compare the test error performance of the OBPM algorithm with other online algorithms: the Perceptron, the voted-Perceptron, and Bagging. We demonstrate that the combinationof the voted-Perceptron algorithm and the OBPM algorithm, called voted-OBPM algorithm has better test error performance than the voted-Perceptron and Bagging algorithms. We investigate the use of various online voting methods against the problem of ranking, and the problem of collaborative filtering of instances. We look at the application of online Bagging and OBPM algorithms to the telecommunications problem of channel equalization. We show that both online methods were successful at reducing the effect on the test error of label flipping and additive noise.¶ Secondly, we introduce a new mixture of experts algorithm, the fixed-share hierarchy (FSH) algorithm. The FSH algorithm is able to track the mixture of experts when the switching rate between the best experts may not be constant. We study the theoretical aspects of the FSH and the practical application of it to adaptive equalization. Using simulations we show that the FSH algorithm is able to track the best expert, or mixture of experts, in both the case where the switching rate is constant and the case where the switching rate is time varying.