No description available.

Dissertation thesis is focused on real digital signal processing by quadrature mirror digital filter banks. In the first part a dual channel quadrature mirror digital filter bank is described briefly. Mainly transfer functions for distorted subband signals are described. In the next part generalized sigma-delta modulator and its linear model are described. Subsequently the generalized sigma-delta modulator is used in decomposition part of quadrature mirror digital filter bank. Designed structure is analyzed. Two design method of transfer functions are designed for the structure on the basis of analysis results. The first method is suitable for hand-made design by intuitive distribution of zeros and poles. The second method is more suitable for computer design. It is iterative method based on correlation. Transfer functions for quadrature mirror digital filter bank with sigma-delta modulators design examples are also part of thesis. Application of designed structure of quadrature mirror digital filter bank can lead to bigger compression ration in lossy data compression.

Résumé : Dans ce projet, un décodage neuronal bayésien est effectué sur le colliculus inférieur du cochon d'Inde. Premièrement, On lit les potentiels évoqués grâce aux électrodes et ensuite on en déduit les potentiels d'actions à l'aide de technique de classification des décharges des neurones. Ensuite, un modèle linéaire généralisé (GLM) est entraîné en associant un stimulus acoustique en même temps que les mesures de potentiel qui sont effectuées. Enfin, nous faisons le décodage neuronal de l'activité des neurones en utilisant une méthode d'estimation statistique par maximum à posteriori afin de reconstituer la représentation spectro-temporelle du signal acoustique qui correspond au stimulus acoustique. Dans ce projet, nous étudions l'impact de différents modèles de codage neuronal ainsi que de différentes représentations spectro-temporelles (qu'elles sont supposé représenter le stimulus acoustique équivalent) sur la précision du décodage bayésien de l'activité neuronale enregistrée par le système auditif central. En fait, le modèle va associer une représentation spectro-temporelle équivalente au stimulus acoustique à partir des mesures faites dans le cerveau. Deux modèles de codage sont comparés: un GLM et un modèle bilinéaire généralisé (GBM), chacun avec trois différentes représentations spectro-temporelles des stimuli d'entrée soit un spectrogramme ainsi que deux représentations bio-inspirées: un banc de filtres gammatones et un spikegramme. Les paramètres des GLM et GBM, soit le champ récepteur spectro-temporel, le filtre post décharge et l'entrée non linéaire (seulement pour le GBM) sont adaptés en utilisant un algorithme d'optimisation par maximum de vraisemblance (ML). Le rapport signal sur bruit entre la représentation reconstruite et la représentation originale est utilisé pour évaluer le décodage, c'est-à-dire la précision de la reconstruction. Nous montrons expérimentalement que la précision de la reconstruction est meilleure avec une représentation par spikegramme qu'avec une représentation par spectrogramme et, en outre, que l'utilisation d'un GBM au lieu d'un GLM augmente la précision de la reconstruction. En fait, nos résultats montrent que le rapport signal à bruit de la reconstruction d'un spikegramme avec le modèle GBM est supérieur de 3.3 dB au rapport signal à bruit de la reconstruction d'un spectrogramme avec le modèle GLM. / Abstract : In this project, Bayesian neural decoding is performed on the neural activity recorded from the inferior colliculus of the guinea pig following the presentation of a vocalization. In particular, we study the impact of different encoding models on the accuracy of reconstruction of different spectro-temporal representations of the input stimulus. First voltages recorded from the inferior colliculus of the guinea pig are read and the spike trains are obtained. Then, we fit an encoding model to the stimulus and associated spike trains. Finally, we do neural decoding on the pairs of stimuli and neural activities using the maximum a posteriori optimization method to obtain the reconstructed spectro-temporal representation of the signal. Two encoding models, a generalized linear model (GLM) and a generalized bilinear model (GBM), are compared along with three different spectro-temporal representations of the input stimuli: a spectrogram and two bio-inspired representations, i.e. a gammatone filter bank (GFB) and a spikegram. The parameters of the GLM and GBM including spectro-temporal receptive field, post spike filter and input non linearity (only for the GBM) are fitted using the maximum likelihood optimization (ML) algorithm. Signal to noise ratios between the reconstructed and original representations are used to evaluate the decoding, or reconstruction accuracy. We experimentally show that the reconstruction accuracy is better with the spikegram representation than with the spectrogram and GFB representation. Furthermore, using a GBM instead of a GLM significantly increases the reconstruction accuracy. In fact, our results show that the spikegram reconstruction accuracy with a GBM fitting yields an SNR that is 3.3 dB better than when using the standard decoding approach of reconstructing a spectrogram with GLM fitting.

Spacecraft missions such as spacecraft docking and formation flying require high precision relative position and attitude data. Although Global Positioining Systems can provide this capability near the earth, deep space missions require the use of alternative technologies. One such technology is the vision-based navigation (VISNAV) sensor system developed at Texas A&M University. VISNAV comprises an electro-optical sensor combined with light sources or beacons. This patented sensor has an analog detector in the focal plane with a rise time of a few microseconds. Accuracies better than one part in 2000 of the field of view have been obtained. This research presents a new approach involving simultaneous activation of beacons with frequency division multiplexing as part of the VISNAV sensor system. In addition, it discusses the synchronous demodulation process using digital heterodyning and decimating filter banks on a low-power fixed point DSP, which improves the accuracy of the sensor measurements and the reliability of the system. This research also presents an optimal and computationally efficient six-degree-of-freedom estimation algorithm using a new measurement model based on the attitude representation of Modified Rodrigues Parameters.

Fixed Broadband Wireless Access (FBWA) is a technology aimed at providing high-speed wireless Internet access, over a wide area, from devices such as personal computers and laptops. FBWA channels are defined in the range of 1-20 MHz which makes the RF front end (RFE) design extremely challenging. In its pursuit to standardize the Broadband Wireless Access (BWA) technologies, IEEE working group 802.16 for Broadband Wireless Access has released the fixed BWA standard IEEE 802.16 – 2004 in 2004. This standard is further backed by a consortium, of leading wireless vendors, chip manufacturers and service providers, officially known as Wireless Interoperability for Microwave Access (WiMAX). In general, any wireless base station (BS), supporting a number of contiguous Frequency Division Multiplexed (FDM) channels has to incorporate an RF front end (RFE) for each RF channel. The precise job of the RFE is to filter the desired channel from a group of RF channels, digitize it and present it to the subsequent baseband system at the proper sampling rate. The system essentially has a bandpass filter (BPF) tuned to the channel of interest followed by a multiplier which brings the channel to a suitable intermediate frequency (IF). The IF output is digitized by an ADC and then brought to the baseband by an appropriate digital multiplier. The baseband samples, thus generated, are at the ADC sampling rate which is significantly higher than the target sampling rate, which is defined by the wireless protocol in use. As a result a sampling rate conversion (SRC) is performed on these baseband samples to bring the channel back to the target sampling rate. Since the input sampling rate need not be an integer multiple of the target sampling rate, Fractional SRC (FSRC) is required in most of the cases. Instead of using a separate ADC and IF section for each individual channels, most systems use a common IF section, followed by a wideband ADC, which operates over a wide frequency band containing a group of contiguous FDM channels. In this case a channelizer is employed to digitally extract the individual channels from the digital IF samples. We formally call this system a receiver channelizer. Such an implementation presents considerable challenge in terms of the computational requirement and of course the cost of the BS. The computational complexity further goes up for FBWA system where channel bandwidth is in the order of several MHz. Though such a system has been analyzed for narrow band wireless systems like GSM, to the best of our knowledge no analysis seems to have been carried out for a wideband system such as WiMAX. In this work, we focus on design of a receiver channelizer for WiMAX BS, which can simultaneously extract a group of contiguous FDM RF channels supported by the BS. The main goal is to obtain a simple, low cost channelizer architecture, which can be implemented in an FPGA. There are a number of techniques available in the literature, from Direct Digital Conversion to Polyphase FFT Filter Banks (PFFB), which can do the job of channelization. But each of them operates with certain constraints and, as a result, suits best to a particular application. Further all of these techniques are generic in nature, in the sense that their structure is independent of any particular standard. With regard to computational requirement of these techniques, PFFB is the best, with respect to the number of complex multiplications required for its implementation. But it needs two very stringent conditions to be satisfied, viz. the number of channels to be extracted is equal to the decimation factor and the sampling rate is a power of 2 times baseband bandwidth. Clearly these conditions may not be satisfied by different wireless communication standards, and in fact, this is not satisfied by the WiMAX standard. This gives us the motivation to analyze the receiver channelizer for WiMAX BS and to find an efficient and low cost architecture of the same. We demonstrate that even though the conditions required by PFFB are not satisfied by the WiMAX standard, we can modify the overall architecture to include the PFFB structure. This is achieved by dividing the receiver channelizer into two blocks. The first block uses the PFFB structure to separate the desired number of channels from the input samples. This process also achieves an integer SRC by a factor that is equal to the number of channels being extracted. This block generates baseband outputs whose sampling rates are related to their target sampling rate by a fractional multiplication factor. In order to bring the channels to their target sampling rate, each output from the PFFB block is fed to a FSRC block, whose job is to use an efficient FSRC algorithm to generate the samples at the target sampling rate. We show that the computational complexity, as compared to the direct implementation, is reduced by a factor, which is approximately equal to the square of the number of channels. After mathematically formulating the receiver channelizer for WiMAX BS, we perform the simulation of the system using a software tool. There are two basic motives behind the simulation of the system which has a mathematical model. Firstly, the software simulation will give an idea whether the designed system is physically realizable. Secondly, this will help in designing the logic for different blocks of the system. Once these individual blocks are simulated and tested, they can be smoothly ported onto an FPGA. For simulation purpose, we parameterize the receiver channelizer in such a way that it can be reconfigured for different ADC sampling rates and IF frequencies, by changing the input clock rate. The system is also reconfigurable in terms of the supported channel bandwidth. This is achieved by storing all the filter coefficients pertaining to each channel type, and loading the required coefficients into the computational engine. Using this methodology we simulate the system for three different IF frequencies (and the corresponding ADC sampling rates) and three different channel types, thus leading to nine different system configurations. The simulation results are in agreement with the mathematical model of the system. Further, we also discuss some important implementation issues for the reconfigurable receiver channelizer. We estimate the memory requirement for implementing the system in an FPGA. The implementation delay is estimated in terms of number of samples. The thesis is organized in five chapters. Chapter 1 gives a brief introduction about the WiMAX system and different existing channelization architecture followed by the outline of the proposed receiver channelizer. In chapter 2, we analyze the proposed receiver channelizer for WiMAX BS and evaluate its computational requirements. Chapter 3 outlines the procedure to generate the WiMAX test signal and specification of the all the filters used in the system. It also lists the simulation parameters and records the results of the simulation. Chapter 4 presents the details of a possible FPGA implementation. We present the concluding remarks and future research directions in the final chapter.

Speech segmentation is the process of identifying the boundaries between words, syllables or phones in the recorded waveforms of spoken natural languages. The lowest level of speech segmentation is the breakup and classification of the sound signal into a string of phones. The difficulty of this problem is compounded by the phenomenon of co-articulation of speech sounds. The classical solution to this problem is to manually label and segment spectrograms. In the first step of this two step process, a trained person listens to a speech signal, recognizes the word and phone sequence, and roughly determines the position of each phonetic boundary. The second step involves examining several features of the speech signal to place a boundary mark at the point where these features best satisfy a certain set of conditions specific for that kind of phonetic boundary. Manual segmentation of speech into phones is a highly time-consuming and painstaking process. Required for a variety of applications, such as acoustic analysis, or building speech synthesis databases for high-quality speech output systems, the time required to carry out this process for even relatively small speech databases can rapidly accumulate to prohibitive levels. This calls for automating the segmentation process. The state-of-art segmentation techniques use Hidden Markov Models (HMM) for phone states. They give an average accuracy of over 95% within 20 ms of manually obtained boundaries. However, HMM based methods require large training data for good performance. Another major disadvantage of such speech recognition based segmentation techniques is that they cannot handle very long utterances, Which are necessary for prosody modeling in speech synthesis applications. Development of Text to Speech (TTS) systems in Indian languages has been difficult till date owing to the non-availability of sizeable segmented speech databases of good quality. Further, no prosody models exist for most of the Indian languages. Therefore, long utterances (at the paragraph level and monologues) have been recorded, as part of this work, for creating the databases. This thesis aims at automating segmentation of very long speech sentences recorded for the application of corpus-based TTS synthesis for multiple Indian languages. In this explicit segmentation problem, we need to force align boundaries in any utterance from its known phonetic transcription. The major disadvantage of forcing boundary alignments on the entire speech waveform of a long utterance is the accumulation of boundary errors. To overcome this, we force boundaries between 2 known phones (here, 2 successive stop consonants are chosen) at a time. Here, the approach used is silence detection as a marker for stop consonants. This method gives around 89% (for Hindi database) accuracy and is language independent and training free. These stop consonants act as anchor points for the next stage. Two methods for explicit segmentation have been proposed. Both the methods rely on the accuracy of the above stop consonant detection stage. Another common stage is the recently proposed implicit method which uses Bach scale filter bank to obtain the feature vectors. The Euclidean Distance of the Mean of the Logarithm (EDML) of these feature vectors shows peaks at the point where the spectrum changes. The method performs with an accuracy of 87% within 20 ms of manually obtained boundaries and also achieves a low deletion and insertion rate of 3.2% and 21.4% respectively, for 100 sentences of Hindi database. The first method is a three stage approach. The first is the stop consonant detection stage followed by the next, which uses Quatieri’s sinusoidal model to classify sounds as voiced/unvoiced within 2 successive stop consonants. The final stage uses the EDML function of Bach scale feature vectors to further obtain boundaries within the voiced and unvoiced regions. It gives a Frame Error Rate (FER) of 26.1% for Hindi database. The second method proposed uses duration statistics of the phones of the language. It again uses the EDML function of Bach scale filter bank to obtain the peaks at the phone transitions and uses the duration statistics to assign probability to each peak being a boundary. In this method, the FER performance improves to 22.8% for the Hindi database. Both the methods are equally promising for the fact that they give low frame error rates. Results show that the second method outperforms the first, because it incorporates the knowledge of durations. For the proposed approaches to be useful, manual interventions are required at the output of each stage. However, this intervention is less tedious and reduces the time taken to segment each sentence by around 60% as compared to the time taken for manual segmentation. The approaches have been successfully tested on 3 different languages, 100 sentences each -Kannada, Tamil and English (we have used TIMIT database for validating the algorithms). In conclusion, a practical solution to the segmentation problem is proposed. Also, the algorithm being training free, language independent (ES-SABSF method) and speaker independent makes it useful in developing TTS systems for multiple languages reducing the segmentation overhead. This method is currently being used in the lab for segmenting long Kannada utterances, spoken by reading a set of 1115 phonetically rich sentences.

Spacecraft missions such as spacecraft docking and formation flying require high precision relative position and attitude data. Although Global Positioining Systems can provide this capability near the earth, deep space missions require the use of alternative technologies. One such technology is the vision-based navigation (VISNAV) sensor system developed at Texas A&M University. VISNAV comprises an electro-optical sensor combined with light sources or beacons. This patented sensor has an analog detector in the focal plane with a rise time of a few microseconds. Accuracies better than one part in 2000 of the field of view have been obtained. This research presents a new approach involving simultaneous activation of beacons with frequency division multiplexing as part of the VISNAV sensor system. In addition, it discusses the synchronous demodulation process using digital heterodyning and decimating filter banks on a low-power fixed point DSP, which improves the accuracy of the sensor measurements and the reliability of the system. This research also presents an optimal and computationally efficient six-degree-of-freedom estimation algorithm using a new measurement model based on the attitude representation of Modified Rodrigues Parameters.

Cette thèse a été effectuée dans le cadre du projet européen FP7 EMPHATIC (Enhanced Multicarrier Techniques for Professional Ad-Hoc and Cell-Based Communications). Plusieurs universités européennes et deux partenaires industriels: THALES Communications Security et CASSIDIAN ont participé à ce projet. L'objectif de ce projet est de développer, d'évaluer et de démontrer les apports des techniques multi-porteuses avancées, permettant une meilleure utilisation des bandes de fréquences radio existantes en fournissant des services de données à large bande, en coexistence avec les services traditionnels à bande étroite. Le projet porte sur l'application de radiocommunications mobiles professionnelles (Professional Mobile Radio : PMR). L'idée principale de ce projet est d'analyser la viabilité des systèmes à large bande utilisant des bancs de filtres (Filter Bank Multi Carrier : FBMC) conjointement avec une modulation d'amplitude en quadrature avec décalage (Offset Quadrature Amplitude Modulation : OQAM) dans le cadre de la 5ème génération (5G) des systèmes radio-mobiles. La modulation FBMC-OQAM se positionne comme candidate potentielle pour les futurs systèmes de communication. Cette modulation avancée offre de nombreux avantages tels que l’excellente localisation fréquentielle de sa densité spectrale de puissance (DSP), une robustesse au bruit de phase, aux décalages de fréquence ainsi qu’à l’asynchronisme entre les utilisateurs. Ces atouts, la rendent plus attrayant qu’OFDM pour l’application PMR, la radio cognitive (CR) et la 5G. Cependant, comme toute autre technique de modulation muti-porteuses, FBMC-OQAM souffre d’un facteur de crête ou d’un PAPR (pour Peak to Average Power Ratio) élevé. Lorsque l'amplificateur de puissance (AP), utilisé au niveau de l’émetteur, est opéré proche de sa zone non-linéaire (NL), ce qui est le cas dans la pratique, la bonne localisation fréquentielle de la DSP du système FBMC/OQAM est sérieusement compromise, en raison des remontées spectrales. Le premier objectif de cette thèse est de prédire l'étendue des remontées spectrales dans les systèmes FBMC-OQAM, introduites par la non-linéarité AP. Le deuxième objectif de ce travail est de proposer des techniques, pour les systèmes FBMC-OQAM, permettant la réduction du PAPR et la linéarisation de l’AP, afin d'atténuer les effets NL. L’utilisation des cumulants, a permis de prédire les remontées spectrales pour les signaux FBMC-OQAM après amplification NL. En outre, certains algorithmes de réduction du PAPR, basées sur des approches probabilistes et des techniques d'ajout de signaux, ont été proposés. La capacité de coexistence du système à large bande utilisant FBMC-OQAM avec des systèmes PMR à bande étroite en présence de PA a été analysée et il a été démontré que la coexistence est possible, à condition qu'il y est une bonne combinaison entre le recul du signal à l’entrée de l’AP (Input Back-Off : IBO), la réduction du PAPR et la linéarisation de l’AP. Enfin, une nouvelle technique de linéarisation de l’AP a été proposée pour le système FBMC-OQAM. / This thesis is part of the European FP7 EMPHATIC project (Enhanced Multicarrier Techniques for Professional Ad-Hoc and Cell-Based Communications) including various European universities and two main industrial partners: THALES Communications Security and CASSIDIAN. The EMPHATIC objective is to develop, evaluate and demonstrate the capability of enhanced multi-carrier techniques to make better use of the existing radio frequency bands in providing broadband data services in coexistence with narrowband legacy services. The project addresses the Professional Mobile Radio (PMR) application. The main idea is to analyze the viability of broadband systems based on filter-bank multi-carrier (FBMC) clubbed with o ffset quadrature amplitude modulation (OQAM) in the context of the future 5th Generation (5G) radio access technology (RAT). Increasingly, the FBMC-OQAM systems are gaining appeal in the probe for advanced multi-carrier modulation (MCM) waveforms for future communication systems. This advanced modulation scheme o ers numerous advantages such as excellent frequency localization in its power spectral density (PSD), a robustness to phase noise, frequency off sets and also to the multi-user asynchronism; making it more appealing than OFDM for PMR, cognitive radio (CR) and 5G RAT. However, like any other MCM technique, FBMC-OQAM suff ers from high PAPR. When the power amplifi er (PA) non-linearity, which is realistic radio-frequency impairment, is taken into account; the good frequency localization property is severely compromised, due to the spectral regrowth. The first objective of this PhD thesis is, to predict the extent of the spectral regrowth in FBMC-OQAM systems, due to the PA non-linearity. The second objective is to probe techniques for FBMC-OQAM systems, such as PAPR reduction and PA linearization, in order to mitigate the NL eff ects of PA. By cumulant analysis, spectral regrowth prediction has been done for FBMC-OQAM systems. Also, some algorithms for PAPR reduction, which are based on probabilistic approach and adding signal methods, have been proposed. The coexistence capability of the FBMC-OQAM based broadband system with the narrowband PMR systems in the presence of PA has been analyzed and it has been found that coexistence is possible, provided there is a symbiotic combination of PA Input Back-off (IBO), PAPR reduction and PA linearization. Finally, a novel PA linearization technique has been proposed for FBMC-OQAM.

The thesis focuses on the issue of signal processing. Using practical examples, it tries to show the use of individual signal processing operations from a practical point of view. For each of the selected signal processing operations, an application is created in MATLAB, including a graphical interface for easier operation. The division of the thesis is such that each chapter is first analyzed from a theoretical point of view, then it is shown using a practical demonstration of what the operation is used in practice. Individual applications are described here, mainly in terms of how they are handled and their possible results. The results of the practical part are presented in the attachment of the thesis.