Dissertations / Theses on the topic 'Systèmes de réveil acoustiques'
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 49 dissertations / theses for your research on the topic 'Systèmes de réveil acoustiques.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
Lévy, Christophe. "Modèles acoustiques compacts pour les systèmes embarqués." Avignon, 2006. http://www.theses.fr/2006AVIG0143.
Full textThe amount of services offered by the last generation mobile phones has significantly increased compared to previous generations. Nowadays, phones offer new kinds of facilitiessuch as organizers, phone books, e-mail/fax, and games. At the same time, the size of mobile phones has steadily reduced. Both these observations raise an important question: ?How can we use the full facilities of a mobile phone without a large keyboard??. Voice based human-to-computer interfaces supply a friendly solution to this problem but require an embedded speech recognizer. Over the last decade, the performance of Automatic Speech Recognition (ASR) systems has improved and nowadays facilites the implementation of vocal human-to-computer interfaces. Moreover, even if scientific progress could be noticed, the potential gain (in performance) remains limited by computing resources: a relatively modern computer with a lot of memory is generally required. The main problem to embed ASR in a mobile phone is the low level of resources available in this context which classically consists of a 50/100 MHz processor, a 50/100 MHz DSP, and less than 100KB of memory. This thesis focuses on embedded speech recognition in the context of limited resources
Stelzer, Rainer. "Une méthode énergétique pour les systèmes vibro-acoustiques couplés." Phd thesis, INSA de Lyon, 2012. http://tel.archives-ouvertes.fr/tel-00780602.
Full textMinard, Antoine. "Perception et confort acoustiques des systèmes de traitement d'air." Phd thesis, Université de La Rochelle, 2013. http://tel.archives-ouvertes.fr/tel-01066154.
Full textRigobert, Stéphane André. "Modélisation par éléments finis des systèmes élasto-poro-acoustiques couplés." Lyon, INSA, 2001. http://www.theses.fr/2001ISAL0014.
Full textIn this work, hierarchical elements are used to discretize the weak { u,P} formulation of the equations of poroelasticity. This approach is validated on three dimensional configurations through a comparison with linear poroelastic elements. The use of high order polynomials for the basic functions improves the convergence of classical linear poroelastic elements and thus allows for an important reduction of the number of degrees of freedom required for the modeling of the porous domain. Similarly to classical finite elements, the establishment of a convergence criterion is difficult when considering the modeling of poroelastic media with hierarchical elements. Hierarchical poroelastic elements are then coupled to an elastic domain modeled with finite elements or a fluid domain modeled with hierarchical elements, assuming non coincident meshes. Considering these two kinds of coupling, the present approach is validated through a comparison with numerical and experimental approaches. Finally, a simplified model for a porous coated plate is presented. In this approach, the elastic plates together with the solid phase of the porous material are modeled by a multilayered plate element. All the damping phenomena in the porous material are accounted for. This model is first validated in the case of a porous coated plate in free field. Its validity is then investigated when the multilayered structure is coupled to a fluid cavity
Lebreton, Jean Mickaël. "Systèmes et protocoles de télé-réveil appliqués à l’optimisation énergétique des réseaux de capteurs sans fil." Thesis, La Réunion, 2017. http://www.theses.fr/2017LARE0014/document.
Full textNowadays, a significant growth of connected things is observed, exceeding even the worldpopulation. Given the magnitude of this phenomenon, the energy efficiency of communicatingobjects is a crucial issue. Maximizing their lifetime is necessary to ensure a qualityof service. In this regard, the aim of this thesis is to optimize the energy consumption ofwireless communications in a wireless sensor network.The concept of wake-up radio was created a decade ago, which consists of waking up thecommunicating node by a remote radio signal. By default, the node remains in sleepingmode at a very low power consumption. If needed, the node can be woken up on demandthrough a specific radio signal. Thus, the energy consumption of the radio module is greatlyreduced by idle listening to the channel in sleeping mode. However, this wake-up radioprinciple requires the development of new hardware architectures associated with adaptedand innovative communication protocols. Despite recent proposals, the level of technologymaturity has not yet been reached on this subject.Therefore, a wake-up radio system is proposed in this thesis, including a theoreticalcharacterization of its performances. The wake-up receiver consumes 363 nW in sleepingmode and 49.8 µW in receiving mode. Moreover, two new protocols called DoRa and DC-DoRaare proposed with an evaluation of their performances by simulation. The resultsshow that these wake-up radio protocols greatly reduce the energy consumed by the radiomodule, compared to the MAC protocols currently used in wireless sensor networks. Finally,the experimental implementation of the wake-up radio system and protocols enabled thereal performance measurement of our approach in an environment with overhearing andinterference
Trenchant, Vincent. "Discrétisation et commande frontière de systèmes vibro-acoustiques, une approche hamiltonienne à ports." Thesis, Bourgogne Franche-Comté, 2017. http://www.theses.fr/2017UBFCD066/document.
Full textThis thesis deals with the boundary control of an acoustic by a network of co-localised sensors/actuators which constitutes a smart skin. In order to cope with this multiphysical problem, we chose to place our study in the framework of port-Hamiltonian systems, a structured approach based on the representation of energy exchanges between different energy domains between different systems of subsystems. We proposed a port-Hamiltonian model of the wave equation interconnected through its boundary to the distributed actuation system, which corresponds to a 2D formulation of the physical problem. We developed a spatial discretization method based on the use of finite differences on several staggered grids that preserve the port-Hamiltonian structure of the wave equation. This method also permits to easily interconnect the discretized system with other subsystems, which is convenient for instance for control purposes. Its main advantage over other structure preserving methods is its simplicity of implementation which stems from the use of finite differences. In order to control the vibro-acoustic system, we proposed a control law synthesis method for systems governed by two conservation laws in 1D. The originality of this method lies in the fact that it relies on the computation of structural invariants (Casimir functions) exploited in order to modify the structure of the system in closed loop. The conditions of application of these laws on a 2D system are studied and numerical results validate the synthesized control laws
Marian, Vlad. "Transmission d'énergie sans fil : Application au réveil à distance de récepteurs en veille zéro consommation." Phd thesis, Ecole Centrale de Lyon, 2012. http://tel.archives-ouvertes.fr/tel-00787608.
Full textChesneau, Hugo. "Modélisation de la déformation de systèmes diphasiques sous contraintes induites par des ondes électromagnétiques ou acoustiques." Thesis, Bordeaux, 2020. http://www.theses.fr/2020BORD0096.
Full textWe numerically studied the deformation of two-phase fluid systems under acoustical or electromagnetic stress. We developed and used a numerical method based on the boundary element method to solve simultaneously the scalar Helmholtz equation and Stokes equation to consider at the same time the wave’s propagation and the hydrodynamic of the system. On the first hand, we considered the deformation of initially flat interfaces actuated by two kinds of waves’ induced stresses. First, we studied the scattering force, a bulk force induced by the propagation and the scattering of an incident electromagnetic waves into a turbid liquid phase. This force leads to the formation of eddies which in turn induce viscous stresses on the interface. Under high laser power, the viscous stresses can destabilize the interface into liquid microjets. We characterized numerically the shapes and the fluid flow of these microjets. Then, we focused on the acoustic and electromagnetic radiation pressure at the fluid/fluid interface. We showed that the induced deformations can be described by a universal model and these deformations act as liquid waveguides who are matching perfectly the incident wave which has induced them. On the other hand, we studied the deformation of micrometric droplets and liposomes with optical tweezers and optical stretcher. For droplets, we retrieved similar behaviour than for flat interfaces i.e., a coupling between the deformation of the droplets and the propagation of the wave into the latter. Additionally, we numerically characterize for the first time the liposome deformation in an optical stretcher and compared our result with recent experiments carried out in the laboratory. These initial studies are promising for the microrheology of liquid drops or biological systems
Ait, Amrane Omar. "Identification de systèmes à réponse impulsionnelle longue par filtrage adaptatif en fréquence : application à l'annulation d’échos acoustiques." Paris, ENST, 1992. http://www.theses.fr/1992ENST0026.
Full textTayong, Boumda Rostand. "Propriétés acoustiques de systèmes incorporant des plaques micro-perforées et des matériaux absorbants sous forts niveaux d'excitation." Thesis, Dijon, 2010. http://www.theses.fr/2010DIJOS066/document.
Full textThis work deals with the acoustical properties of systems incorporating Micro-Perforated Panels (MPP) and absorbing materials under high level of excitation.In the first chapter, absorbent systems composed of an air-cavity backed MPP are studied at high level of excitations. An analytical model involving two dimensionless parameters and an optimum Mach number is proposed. This model describes the behavior of the maximum of absorption coefficient (absorption coefficient at the resonance) with respect to the Mach number inside the perforations. A formula is proposed that predicts the variations of the absorption peak with the acoustical Mach number.In the second chapter, the holes interaction effects are studied theoretically and experimentally under high levels of excitations. Following an equivalent fluid approach, a model for which the tortuosity is corrected to account for the holes interaction effects coupled to the jet-like effects is developed. Multi-layered absorbents composed of MPP and porous materials are then studied under high level of excitations. The case where the multi-layers are directly attached to a rigid wall and the case where there is an air cavity before the rigid wall are examined. Forchheimer's law is used to model each medium of the multi-layer and the use of the transfer matrix method is made to account for these media.Sound transmission study under high level of excitation is introduced. The perspectives of this work are numerous and promising in the acoustics of transportation systems applications
Zimpfer, Véronique. "Amélioration du traitement numérique des signaux dans les systèmes actifs en protection auditive." Lyon, INSA, 2000. http://theses.insa-lyon.fr/publication/2000ISAL0101/these.pdf.
Full textIn order to realize an active noise reduction (ANR) on an active hearing protector, with a digital filter, it is necessary that the filter works in real time with a broad bandwidth and a large dynamic range. Only the IIR type digital filters can comply with the real time condition. However, the quantification effect of coefficients and signals and the propagation of roundoff noise decrease the filter performance. Different solutions have been adapted in order to minimize the quantification effects. In order to increase the filter dynamics, it has been necessary to create a new algorithm for the IIR type filters. This algorithm is called "adapted algorithm". Its essential part is to minimize the propagation of the roundoff noise to the next sampling of output signals. Hence, it has been possible to realize active hearing protectors using digital filters. The digital filter allows to adjust the open loop gain in order to have the optimum active attenuation. We have confirmed that the hearing protectors using digital filters are as efficient as those using analog filters. By using digital filters, the hearing protectors become more flexible than with analog filters. The use of a digital ANR system allows to easily add subsidiary functionalities to the ANR system, as for example : the insertion of a speech signal, an automatic "go-stop" procedure and an ANR system using two filters
Rybak, Fanny. "Systèmes de codage et de décodage des signaux acoustiques émis pendant la parade de cour de Drosophila Melanogaster." Paris 13, 1999. http://www.theses.fr/1999PA132042.
Full textDeviren, Murat. "Systèmes de reconnaissance de la parole revisités : réseaux bayésiens dynamiques et nouveaux paradigmes." Nancy 1, 2004. http://www.theses.fr/2004NAN10161.
Full textIn this thesis we focus on four principle components of a speech recognition system: acoustic modeling, language modeling, speech feature extraction and noise compensation. We propose novel modeling approaches for acoustic and linguistic modeling within the Bayesian networks formalism. Bayesian networks are a subset of probabilistic graphical models that include the most widely used probability models in speech recognition. Therefore rethinking the modeling problems in this formalism provides new perspectives that were not considered previously. Besides novel modeling approaches we also address new speech feature extraction schemes. Our main motivation in this direction is to seek for robust features that are not bound to be used in classical hidden Markov modeling (HMM) approach. Finally, we address the robustness problem for varying application conditions and propose a novel supervised compensation scheme
Kone, Issiaka. "Résonateurs à ondes acoustiques guidées sur miroir de Bragg." Thesis, Lille 1, 2010. http://www.theses.fr/2010LIL10060/document.
Full textPassive acoustic devices are enabling elements for RF architectures. Aside from the well known surface acoustic wave (SAW) and bulk acoustic wave (BAW) filters used for RF filtering, Lamb wave devices have recently been seen as an alternative to SAW devices currently used also for intermediate frequency filtering. These resonators are compatible with a membrane-based BAW technology (FBAR) and could thus be co-integrated with them with no additional cost. However, STMicroelectronics and CEALeti have focused their developments around BAW resonators using a Bragg mirror for acoustic isolation (SMR). Therefore, the aim of this work is to demonstrate the possibility of fabricating resonators close to Lamb wave resonators, but technologically compatible with a BAW-SMR technology. After reviewing the applications of electromechanical resonators in RF architectures and the proposals for the kind of resonators we are investigating available in the literature, we start with a theoretical investigation of these devices. This enables us to determine basic characteristics of these resonators: electromechanical coupling factors close to Lamb wave devices, but operation at frequencies close to BAW resonators. Then, we describe the design of demonstrators. Finally, we describe a fabrication process, similar to BAW-SMR fabrication process, which has provided to date the first operating devices using this principle. Electrical characterizations show that the fabricated resonators are functional, and we discuss possible improvements.This work led to publications in four conferences, (three international conferences), a paper, a patent and two internal publications at IEMN-STMicroelectronics common laboratory
Touzé, Cyril. "Analyse et modélisation de signaux vibratoires et acoustiques chaotiques : application aux instruments de percussion non-linéaires." Paris 6, 2000. http://www.theses.fr/2000PA066571.
Full textMareschal, Olivier. "Étude d’un résonateur piézoélectrique à ondes acoustiques de volume en technologie film mince." Thesis, Paris Est, 2011. http://www.theses.fr/2011PEST1035/document.
Full textThe studied resonator is part of an industrial project carried by NXP Semiconductors. The objective is the realization of a integrable RF MEMS resonator in order to replace quartz in some applications. The compatibility of the manufacturing process with the technologies used by the company and low cost production represent the main challenges of the project. The resonator TFEAR (Thin Film Elongation Acoustic Resonator) is a bar, consisting of a superposition of thin film type Metal/AlN/metal. The piezoelectric properties of aluminum nitride (AlN) are exploited : the application of an alternating electric field, parallel to the thickness of the bar, resulting in propagation of acoustic waves along its length. The sizes of the manufactured resonators correspond to resonant frequencies between 10MHz to 50 MHz. This thesis focuses on modeling and electrical characterization of the TFEAR resonator. The models are developed by 3D numerical simulations and by 1D analytical calculations. The electrical behavior of TFEAR is described by an equivalent circuit which elements are expressed in terms of physical parameters and losses of the constituent materials. A quality factor of 2250 on a 25.79MHz resonant TFEAR which motional resistance is 2.1 kOhms has been noticed. These measurements were completed by the characterization of the physical parameters of the piezoelectric layer. For example, piezoelectric coefficient d33;f values were recorded up to 2.6 pm/V (for a theoretical maximum of 3.93pm/V)
Michaud, Pierre-Yohan. "Distorsions des systèmes de reproduction musicale : Protocole de caractérisation perceptive." Phd thesis, Aix-Marseille Université, 2012. http://tel.archives-ouvertes.fr/tel-00806288.
Full textCuré, Charlotte. "Etude des systèmes de communication acoustique des puffins du bassin méditerranéen : puffinus yelkouan, puffinus mauretanicus et calonectris diomedea diomedea." Paris 11, 2008. http://www.theses.fr/2008PA112130.
Full textShearwaters (procellariidae) are burrowing seabirds. They return to their colony at dusk and exhibit a high vocal activity. One single call is used in both sexual and territorial contexts. We conducted a comparative study of the acoustic communication systems of the three shearwater species in the Mediterranean basin (puffinus yelkouan, puffinus mauretanicus and calonectris d. Diomedea). These species showing overlapping nesting areas, we focused on the species specific signature in the call. Considering the fidelilty to the mate and the incubation relays between partners, we also studied the sexual and individual vocal signatures. By performing acoustic analyses, we showed that these signatures were multiparametrics. By playback experiments, we showed that birds of three species were able to vocally identify the gender of their congeners. Replying only to conspecific same-sex calls, birds ensure burrow defence and mate guarding. Concerning the communication strategies between partners, significant differences were observed between species. Contrary to puffinus yelkouan, calonectris d. Diomedea vocally replied to its mate. For the former, an additional channel (olfactory for example) is probably necessary to ensure mate identification. Concerning the species-specific recognition, the differences in the acoustic codes observed between the species appeared all the more significant as the phylogenetic distance was high. Thus, even submitted to similar environmental and social constraints, Mediterranean shearwater species show differences in their communication system
Goryachev, Maxim. "Résonateurs à ondes acoustiques de volume et oscillateurs à température de l'hélium liquide." Phd thesis, Université de Franche-Comté, 2011. http://tel.archives-ouvertes.fr/tel-00651960.
Full textZhou, Huan. "Etude théorique et expérimentale de systèmes à ondes de surface dans des structures multicouches piézomagnétiques pour des applications en contrôle santé intégré de MEMS par imagerie acoustique non linéaire." Phd thesis, Ecole Centrale de Lille, 2014. http://tel.archives-ouvertes.fr/tel-00991915.
Full textBennouna, Saâd. "Caractérisation aéroacoustique d'éléments et associations d'éléments de systèmes de ventilation d'air pour l'automobile." Thesis, Compiègne, 2016. http://www.theses.fr/2016COMP2303/document.
Full textPassenger’s thermal comfort inside car cabin is mainly provided by the heating ventilation and air conditioning system (HVAC). The main part of HVAC modules is placed under the dash board. An HVAC module is a compact system composed of various elements which are subject to airflow. The interaction between airflow and these in-duct elements generates noise inside car cabin. Furthermore, the blower used to blow air inside the cabin must overcome the pressure generated by HVAC elements. Noise is created and its level is linked to flow and pressure. HVAC noise is an important issue for car makers and automotive suppliers wishing to reach passenger’s satisfaction. Furthermore, thermal-engine cars are more and more silent. Also hybrid and electric car sells are expanding around the world. HVAC noise became a main issue for automotive actors. Under CEVAS project, Valeo is aiming to develop a prediction tool to design HVAC systems providing sound quality data. Within CEVAS project the UTC is in charge of performing measurements on academic elements and industrial HVAC components : diaphragm, flaps,…
Ricaud, Fabienne. "Etude de l'identification des sources acoustiques à partir du couplage de la pression en champ proche et de l'organisation instantanée de la zone de mélange de jet." Poitiers, 2003. http://www.theses.fr/2003POIT2274.
Full textRamoul, Hichem. "Inégalités de Carleman pour des systèmes paraboliques et applications aux problèmes inverses et à la contrôlabilité : contribution à la diffraction d'ondes acoustiques dans un demi-plan homogène." Thesis, Aix-Marseille 1, 2011. http://www.theses.fr/2011AIX10025.
Full textIn the first part, we prove Carleman estimates for parabolic systems. In chapter1, we prove stability inequalities for 2 x 2 parabolic system using Carleman estimates with one observation. It is concerns to the identification of the coefficients and initial conditions of the system. The chapter2 is devoted to th Carleman estimates of parabolic systems for which the diffusion coefficients are assumed to be ofclass piecewise C1 or with bounded variations. In the end, we give some applications to the null controllability. The second part is devoted to the study of the scattering problem of acoustics waves in a homogeneous half-plane. It is about a boundary value problem associated to the Helmholtz equation in theupper half-plane with a nonhomogeneous Neumann boundary data. We provide some answers to the question of uniqueness and existence of solutions for some classes of the boundary data
Ramamonjy, Aro. "Développement de nouvelles méthodes de classification/localisation de signaux acoustiques appliquées aux véhicules aériens." Thesis, Paris, CNAM, 2019. http://www.theses.fr/2019CNAM1234/document.
Full textThis thesis deals with the development of a compact microphone array and a dedicated signal processing chain for aerialtarget recognition and direction of arrival (DOA) estimation. The suggested global approach consists in an initial detection ofa potential target, followed by a DOA estimation and tracking process, along with a refined detection, facilitated by adaptivespatial filtering. An original DOA estimation algorithm is proposed. It uses the RANSAC algorithm on real-time time-domainbroadband [100 Hz - 10 kHz] pressure and particle velocity data which are estimated using finite differences and sums ofsignals of microphone pairs with frequency-dependent inter-microphone spacings. The use of higher order finite differences, or variants of the Phase and Amplitude Gradient Estimation (PAGE) method adapted to the designed antenna, can extend its bandwidth at high frequencies. The designed compact microphone array uses 32 digital MEMS microphones, horizontally disposed over an area of 7.5 centimeters. This array geometry is suitable to the implemented algorithms for DOA estimation and spatial filtering. DOA estimation and tracking of a trajectory controlled by a spatialization sphere in the Ambisonic domain have shown an average DOA estimation error of 4 degrees. A database of flying drones acoustic signatures has been set up, with the knowledge of the drone’s position in relation to the microphone array set out by GPS measurements. Adding artificial noise to the data, and selecting acoustic features with evolutionary programming have enabled the detection of an unknown drone in an unknown soundscape within 200 meters with the JRip classifier. In order to facilitate the detection and extend its range, the initial detection stage is preceded by differential beamforming in four main directions (north, south, east, west), and the refined detection stage is preceded by MVDR beamforming informed by the target’s DOA
Vanotti, Meddy. "Développement d'un système de détection en milieux gazeux d'espèces à risque pour le contrôle environnemental (application au monoxyde de carbone et à l'hydrogène) : Composants et systèmes micro-acoustiques." Thesis, Besançon, 2015. http://www.theses.fr/2015BESA2022/document.
Full textThe detection of hazardous gas is a topical issue for the protection of persons. Besides, it represents a challenge linked to the storage of renewable energy. Simulation tools developed within the Time and Frequency Department attached to the FEMTO-ST Institute together with technological facilities available at MIMENTO center have enabled the development of SAW sensors providing answers to these issues. These Love wave’sbased sensors properties have enabled the detection of carbon monoxide in the ppm range. Similarly, hydrogen concentrations of the percent order has been measured by mean of Rayleigh wave’s based sensors. The efforts to optimize electro-acoustic devices have led to achieve delay lines built on quartz with insertion losses of 16 dB.Usually around 25 dB to 30 dB, the reduction of the insertion losses improves the potential of these sensorsin terms of autonomy and surface functionalization. Knowledge of the physical phenomena governing theoperation of these sensors represent the basis of their future development. Thus, different characterization and analysis techniques available in our institute have been carried out to reveal these phenomena. From there, the functionalization of sensor’s sensitives surfaces with metallic alloys and implementation of a chip separation method limiting the disturbance of the direct signal of the electro-acoustic devices, have been explored to improve the performance of the sensors. Based on the experimental results obtained in this thesis, the potential of elastic guided wave’s sensors applied to the detection of chemical quantities in gas phase can be established.In continuation of this study, two projects (P-AIR and SMARTY) dedicated to the control of the urban air quality have already been engaged
Li, Mingdong. "Etude et réalisation de sources stables microondes à base de résonateur BAW pour micro-horloge atomique au Césium." Thesis, Lille 1, 2011. http://www.theses.fr/2011LIL10006/document.
Full textThe objective of this thesis is to study different topologies of low phase noise oscillator as well as a frequency synthesizer at 4.6 GHz for the application of Cs vapor microcell atomic clock. We have studied 7 topologies of low-phase-noise oscillators combined with BAW resonator and 0.25μm SiGe:C BiCMOS Technology from ST Microroelectronics at 2.1 GHz, which present the stat of the art performances. An experimental comparison of these different topologies confirmed the theoretical results. In particular, experimental results showed that the differential topology has better performance in phase noise and is less sensitive to the noise from power supply compared to the single-ended topologies. Finally, a Double-differential-Colpitts topology was proposed, because it was theoretically possible to obtain performance beyond the state of the art in phase noise and in size. Different architectures of frequency synthesizer at 4.6 GHz were studied. We have realized a frequency synthesizer in which a DDS controlled by a FPGA was used. To get good performance in terms of purity of the spectrum, different commercial circuits constituting the frequency synthesis were finely characterized. Finally, the proposed architecture has been used within a completed Cs vapor microcell atomic clock. We got a signal locked with a stability of 2.5 10 – 9 at 1second
Li, Sizhe. "Optimisation technologique d'un laboratoire sur puce intégrant des fonctions acoustiques hautes fréquences : premières applications à l'actionnement en canal microfluidique." Thesis, Valenciennes, 2016. http://www.theses.fr/2016VALE0014/document.
Full textThe interest of ultrasounds for media characterization or for actuation when using more power is well known. Nevertheless, the integration of these acoustic functions in silicon based Lab-on-chips requires specific technological developments. The possibility to use high frequency bulk acoustic waves in this kind of systems for characterization or detection has been presented previously in another PhD work. The main objective of this work was to optimize acoustic energy transfer to a microfluidic channel in a frequency range between 500 MHz and 1000 MHz. To do that, the main technological developments achieved among others concern the coating of the guiding mirrors to avoid acoustic mode conversion and ZnO thick films sputtering for the fabrication of piezoelectric transducers. The developed system has been used for particles detection or concentration evaluation. Moreover, a first evaluation of fluids/particles actuation was demonstrated, along with temperature evaluation using ultrasound were achieved in microfluidic channels
Launay, Thierry. "Prise en compte en temps réel d'événements à apparition aléatoire dans le raisonnement d'un système expert en ligne : application à la surveillance acoustique des générateurs de vapeur." Compiègne, 1989. http://www.theses.fr/1989COMPD163.
Full textDevaux, Cédric. "Modélisation du comportement vibratoire des structures par des méthodes énergétiques : formulation moyennée spatialement pour des systèmes unidimensionnels." Phd thesis, Université du Maine, 2006. http://tel.archives-ouvertes.fr/tel-00132382.
Full textDans un premier temps, le concept de superposition quadratique est présenté : si les variables linéaires telles que le déplacement associé à l'onde sont la somme de n composantes différentes, toute variable quadratique telle que l'intensité ou les densités d'énergie peut être présentée sous la forme de n² termes différents.
Ceci est illustré notamment dans le cas bidimensionnel de la superposition de deux ondes planes.
Dans un second temps, le cas unidimensionnel de deux ondes planes contre-propagatives est étudié car il fournit les variations des champs énergétiques à deux échelles bien distinctes.
A petite échelle, les variations des champs énergétiques représentent la
structure locale des interférences définies par un nombre d'onde purement réel.
A grande échelle, les variations des champs énergétiques représentent les transferts énergétiques globaux dus à la dissipation et définies par un nombre d'onde purement imaginaire.
La partie suivante est quant à elle consacrée aux vibrations de plaques.
Différents types d'ondes sont considérés: ondes quasi-longitudinales, ondes de cisaillement et ondes de flexion.
Dans les cas unidimensionnels (plaques semi-infinies), l'analyse pour les ondes quasi-longitudinales et les ondes de cisaillement s'avère similaire à celle présentée précédemment. En revanche le cas des ondes de flexion s'avère plus compliqué en raison de la présence de composantes évanescentes dans le champ de déplacement, lesquelles multiplient d'autant le nombre de composantes des variables énergétiques.
Une formulation quadratique équivalente à celle en déplacement a néanmoins pu être obtenue pour les ondes de flexion unidimensionnelles.
Enfin la dernière partie montre tout d'abord comment une formulation quadratique moyenne peut être développée dans le cas d'ondes planes unidimensionnelles, l'opération de moyennage permettant de s'affranchir des composantes à petite échelle spatiale des variables quadratiques pseudo-périodiques.
Une équation différentielle est obtenue pour l'intensité complexe, les densités d'énergie pouvant être tirées de cette variable.
Les conditions limites énergétiques tenant compte des composantes active et réactive de l'intensité sont ensuite calculées, pour des jonctions passives ou actives. Les cas de jonctions passives font intervenir des conditions mixtes analogues aux conditions d'impédance d'une formulation en déplacement.
Le cas des jonctions actives fait quant à lui intervenir non seulement des impédances mais également la densité de puissance injectée dans la discontinuité d'intensité moyennée. Cette formulation quadratique moyenne peut alors être appliquée au domaine des moyennes fréquences.
Arab, Mohamed. "Analyse des systèmes pétroliers de l'offshore algérien oriental : quantification, modélisation stratigraphique et thermique." Thesis, Brest, 2016. http://www.theses.fr/2016BRES0130/document.
Full textThe Algerian margin is a complex domain, limited to the south by the Maghrebian internal zones (AlKaPeCa) that overthrust the external Tellian zones southward. The offshore part of this system constitutes a back-arc basin, where a mio-pliocene sedimentary series were deposited over a substratum of laterally variable origin and nature. The goal of the present work consisted in analyzing and modelling the petroleum system at work and calculating hydrocarbon volumes. Before proceeding to different numerical basin, stratigraphic and thermal modelling, a conceptual geological model is required and was performed based on field geological studies and stratigraphic and structural interpretations of the seismic profiles. The extension of the continental crust beyond the foot of the margin, allows determining a chronostratigraphic model by extrapolation of the outcroping oligo-miocene formations onshore to the acoustic pre-messinian units defined in the offshore basin. Besides, a tectonic analysis in both onshore and offshore domains gave rise to three main steps of evolution: (1) syn-rift phase where the basin was opened by transtension since Late Oligocene- Aquitanian, (2) post-rift phase in which intra-arc basins were developed, AlKaPeCa and Africa docked and magmatism activity took place, (3) inversion phase where the margin became active mainly since Quaternary times. After defining the different petroleum system elements taking into account variable uncertainties, the 2D/3D petroleum system model depicts possibilities of oil and gas accumulations mainly close to the margin, between 20 to 65 km from the coastline to the north with a maximum range of 70 km in the Bejaia Gulf
Liu, Yuxin. "Etude de cristaux phononiques à base des matériaux micro/nano structurés pour la manipulation des ondes de Love." Thesis, Ecole centrale de Lille, 2019. http://www.theses.fr/2019ECLI0007.
Full textThe control of the propagation of elastic waves relies mainly on the design of artificial medium based on structured materials to obtain an advanced engineering of the dispersion of the propagation. During the thesis, the dispersion of the shear horizontal polarised guided mode (Love mode) in the bi-layer SiO2/Quartz (90ST-cut) structure was numerically investigated and the resulting applications explored. The properties of phononic crystals (PnCs) based on micro-machined holes in the SiO2 layer, and the interaction of this mode with pillars deposited on the surface of this guiding layer, have been studied. In the case of holey PnCs we have shown that it is possible to open band gaps, this property has been exploited for the design of a cavity resonator. The performances of this resonator are studied according to the geometrical parameters characterizing it. It is also proposed to study the interaction of the cavity resonator’s modes with the resonant modes of pillars deposited on the surface of the cavity. This has the effect of a better confinement of the modes and thus a drastic improvement of the quality factor. We also investigated the interaction between the Love mode and metasurfaces based on pillars deposited on the surface of SiO2. The couplings between pillars of identical or not geometries gave rise to various phenomena like acoustic analogue of Autler-Townes Splitting, Fabry-Perot resonance, acoustically induced transparency and Fano resonance. The results presented in this study could be used for potential applications such as signal processing, wave control, metamaterials and biosensors
Arab, Mohamed. "Analyse des systèmes pétroliers de l'offshore algérien oriental : quantification, modélisation stratigraphique et thermique." Electronic Thesis or Diss., Brest, 2016. http://www.theses.fr/2016BRES0130.
Full textThe Algerian margin is a complex domain, limited to the south by the Maghrebian internal zones (AlKaPeCa) that overthrust the external Tellian zones southward. The offshore part of this system constitutes a back-arc basin, where a mio-pliocene sedimentary series were deposited over a substratum of laterally variable origin and nature. The goal of the present work consisted in analyzing and modelling the petroleum system at work and calculating hydrocarbon volumes. Before proceeding to different numerical basin, stratigraphic and thermal modelling, a conceptual geological model is required and was performed based on field geological studies and stratigraphic and structural interpretations of the seismic profiles. The extension of the continental crust beyond the foot of the margin, allows determining a chronostratigraphic model by extrapolation of the outcroping oligo-miocene formations onshore to the acoustic pre-messinian units defined in the offshore basin. Besides, a tectonic analysis in both onshore and offshore domains gave rise to three main steps of evolution: (1) syn-rift phase where the basin was opened by transtension since Late Oligocene- Aquitanian, (2) post-rift phase in which intra-arc basins were developed, AlKaPeCa and Africa docked and magmatism activity took place, (3) inversion phase where the margin became active mainly since Quaternary times. After defining the different petroleum system elements taking into account variable uncertainties, the 2D/3D petroleum system model depicts possibilities of oil and gas accumulations mainly close to the margin, between 20 to 65 km from the coastline to the north with a maximum range of 70 km in the Bejaia Gulf
Taillan, Christophe. "Structuration dynamique du substrat lors de la croissance épitaxiale." Toulouse 3, 2012. http://thesesups.ups-tlse.fr/1778/.
Full textIn semiconductor or metallic components, reduction of active zones to nanometric dimensions leads to an improvement of their performance. Thus, quantum confinement properties of semiconductor quantum dots and arrays of have opened new solutions for the realization of optoelectronic and photonic components. Likewise for metal components, reducing active magnetic areas can significantly improve the storage of information. Quality of these new components needs a fine control of spatial distribution and size distribution of the structures thus created. The process normally used, for the realization of these nanostructures, is the epitaxial growth. A major current objective is to use self-organization properties of atoms or island on the substrate to avoid the use of masking or costly lithography techniques. In this context, different techniques, such as Stranski-Krastanov growth and buried dislocation networks, use elastic properties of materials. This thesis presents the theoretical study of an alternative approach that will allow fine control of the position and size distribution of nanostructures. Making an analogy with the formation of figures Chaldni at the macroscopic scale, we propose to study the effect of a standing acoustic vibration to the surface of a substrate during the epitaxial growth. The diffusion of an adatom on the free surface of a crystalline substrate submitted to a one-dimensional standing acoustic vibration is studied by molecular dynamics simulations. The probability of finding the adatom in the vicinity of the antinodes of transverse displacement of the surface is significantly higher than that to find in the vicinity of a node. The effect of the wave is studied as a function of the amplitude, frequency and direction of the wave, the surface of the substrate and the temperature. In the second part, one-dimensional model of diffusion of an adatom on a substrate submitted to a standing wave is solved analytically. This study allows to find and understand the results of the simulations. A third part presents the results obtained for the diffusion of clusters on a surface. The results of the very small clusters are similar to those obtained for atomic diffusion. Preliminary calculations on larger clusters obtained by dewetting of a surface show that they can position themselves preferentially in the vicinity of the antinodes or nodes of the wave according to the solid or liquid phase of the cluster
Georgel, Vincent. "Étude de l'influence de l'assemblage sur le comportement des composants électromécaniques intégrés dans des systèmes radiofréquences." Phd thesis, Université Paris-Est, 2008. http://tel.archives-ouvertes.fr/tel-00432057.
Full textMignot, Rémi. "Réalisation en guides d'ondes numériques stables d'un modèle acoustique réaliste pour la simulation en temps-réel d'instruments à vent." Phd thesis, Télécom ParisTech, 2009. http://tel.archives-ouvertes.fr/tel-00456997.
Full textMaréchal, Romain. "Modélisation par méthode mixte analytique-numérique de la matrice de diffusion d'un conduit traité de longueur finie avec tubes Herschel-Quincke intégrés : application à la réduction du bruit de soufflante." Compiègne, 2011. http://www.theses.fr/2011COMP1988.
Full textIn aeronautic industry, the efficiency of acoustic liners in bypass and inter-stage regions of the turbofan engine is reached the limit, especially at the Blade Passage Frequencies (BPF), due to the interaction between the fan blades and the struts. Then, pure tones are produced in the noise spectrum for all flight phases, in addition of acoustic liners, Herschel-Quincke (HQ) tubes provide an alternative to improve the attenuation at the BPF. These tubes are placed in the periphery of the main duct and produce phase differences between acoustic waves in the main duct and in the HO tubes in order to attenuate the transmitted acoustic power. A mixed numerical-analytical model was developed for the fast computation of the scattering matrix of a ring of HO tubes integrated in a finite length line wall circular duct. This technic takes into account the exact shape of HO tube and the possibility of a non uniform acoustic velocity at the interfaces between the main duct and the HO tube. This model also allows a study of HQ tube parameters with a time swing compared to a standard element method. Finally, the mixed model has no frequency limit
Issenmann, Bruno. "Déformations d'interfaces fluides par la pression de radiation acoustique." Bordeaux 1, 2007. http://www.theses.fr/2007BOR13505.
Full textVanwynsberghe, Charles. "Réseaux à grand nombre de microphones : applicabilité et mise en œuvre." Thesis, Paris 6, 2016. http://www.theses.fr/2016PA066474/document.
Full textRecently, digital MEMS microphones came out and have opened new perspectives. One of them is the design of large-aperture and massively multichannel acoustical acquisition systems. Such systems meet good requirements for efficient source localization. However, new problems arise. First, an important data flow comes from the array, and must be processed fast enough. Second, if the large array is set up in situ, retrieving the position of numerous microphones becomes a challenging task. This thesis proposes methods addressing these two problems. The first part exhibits the description of the acquisition system, which has been developed during the thesis. First, we show that MEMS microphone characteristics are suitable for array processing applications. Then, real-time processing of channel signals is achieved by a parallel GPU implementation. This strategy is one solution to the heavy data flow processing issue. In this way, a real-time acoustic imaging tool was developed, and enables a dynamic wide-band diagnosis, for an arbitrary duration.The second part presents several robust geometric calibration methods: they retrieve microphone positions, based only on the array acoustic signals. Indeed, in real-life conditions, the state of the art methods are inefficient with large arrays. This thesis proposes techniques that guarantee the robustness of the calibration process. The proposed methods allow calibration in the different existing soundscapes, from free field to reverberant field. Various experimental scenarios prove the efficiency of the methods
Tahon, Marie. "Analyse acoustique de la voix émotionnelle de locuteurs lors d’une interaction humain-robot." Thesis, Paris 11, 2012. http://www.theses.fr/2012PA112275/document.
Full textThis thesis deals with emotional voices during a human-robot interaction. In a natural interaction, we define at least, four kinds of variabilities: environment (room, microphone); speaker, its physic characteristics (gender, age, voice type) and personality; emotional states; and finally the kind of interaction (game scenario, emergency, everyday life). From audio signals collected in different conditions, we tried to find out, with acoustic features, to overlap speaker and his emotional state characterisation taking into account these variabilities.To find which features are essential and which are to avoid is hard challenge because it needs to work with a high number of variabilities and then to have riche and diverse data to our disposal. The main results are about the collection and the annotation of natural emotional corpora that have been recorded with different kinds of speakers (children, adults, elderly people) in various environments, and about how reliable are acoustic features across the four variabilities. This analysis led to two interesting aspects: the audio characterisation of a corpus and the drawing of a black list of features which vary a lot. Emotions are ust a part of paralinguistic features that are supported by the audio channel, other paralinguistic features have been studied such as personality and stress in the voice. We have also built automatic emotion recognition and speaker characterisation module that we have tested during realistic interactions. An ethic discussion have been driven on our work
Jendrzejczak, Christophe. "Développement de techniques de séparation et d’identification de capteurs passifs SAW." Thesis, Université Côte d'Azur, 2021. http://www.theses.fr/2021COAZ4119.
Full textDue to their purely passive nature, sensors using Surface Acoustic Wave (SAW) technology are of great interest in severe environments (strong electromagnetic fields, high temperature ...). These sensors are mainly based on the paralleling of resonators whose frequency will vary depending on the temperature, each sensor occupying a defined frequency band (sub-band and frequency multiplexing). One of the current limitations is the bandwidth of the ISM bands in Europe, which allows only a small number of sub-bands and therefore sensors to be managed. Two methods have been studied to solve the problems of identification and separation of measurements from SAW sensors.The first one, known as the radiation null method directly applicable in the case of two sensors, is based on the use of a 434 MHz reader which has two out-of-phase antenna outputs and consists in adjusting the power and the phase of the two emitted signals. Each of the antennas create a null radiation toward one of the two sensors. This method has been first validated in simulations and then experimentally.The second method is based on the measurement of the superposition of the temporal responses of the SAW resonators (damped oscillations each characterized by four parameters: amplitude A, phase φ, resonance frequency f and damping σ) at two spatial points and the post-processing of these measurements using the high-resolution techniques introduced by Prony in the early 19th century. This method is advantageous because it enables to operate in the case of N sensors (N greater than or equal to 1) with the use of two antennas only connected to the reader. This method makes it possible, through an adequate sampling of the temporal signal, to construct a system of equations whose resolution leads to the determination of the four parameters A, φ, f and σ for every resonator, the frequency information is the desired parameter. For example, we can extract the temperature in the case of a thermal sensor. The system is fixed; the phase difference of the waves re-emitted by the resonators makes it possible to identify the sensors
Hamdad, Hichem. "Recherche d'une description optimum des sources et systèmes vibroacoustiques pour la simulation du bruit de passage des véhicules automobiles." Thesis, Le Mans, 2018. http://www.theses.fr/2018LEMA1032/document.
Full textCurrently, to put a vehicle on market, car manufacturers must comply to a certification test of exterior noise. The regulation of the United Nations Economic Commission for Europe, ECE R51-03, specifies permissible levels a rolling motor vehicle can emit. This regulation is applied since July 1st, 2016, to replace the old regulation ECE R51-02 (test method change and tightening of permissible levels). The drastic reduction in noise levels will be done in 3 steps: from 74 dB (A) under the old regulation to 68 dB (A) in 2024. Therefore, manufacturers as well as their suppliers will face a great challenge to achieve this goal. The objective of this thesis is to develop an aid to the modeling of the pass-by noise of a vehicle, as called for in regulatory testing. The goal is to predict and evaluate accurately the noise emissions earlier in the vehicle development cycle, i.e. before the industrialization stage. We must then seek a trade-off between accuracy of estimates, sensitivity to parameters, robustness of the method and numerical efficiency
Houdouin, Alexandre. "Vers une paroi acoustique absorbante en technologie MEMS." Thesis, Le Mans, 2014. http://www.theses.fr/2014LEMA1020/document.
Full textThe work presented in this thesis focuses on the development of a sound absorbent thin solution able to absorb sound waves of low frequency (500 - 1500 Hz). Noise is, actually, the primary source of environmental pollution raised by the public. This discomfort requires the establishment of acoustic solutions in order to improve the acoustic comfort. However, under certain conditions, the thickness of absorbent solutions strongly limit their use. Indeed, in general, more frequencies are low more the acoustic solutions used must be thick. The sound absorption noise of the solution presented in this work is based on a network of miniature electrodynamic transducers controlled from appropriate electrical loads connected to the terminals of the transducers. An analytical model of the behavior of sound absorbing wall was developed. This model takes into account the behavior of electrodynamic transducers used and the acoustic coupling between the various sources that are particularly important in the area of low frequencies. This model has been validated by two means : i) finite element modeling and ii) measuring the absorption of acoustic prototypes. Two types of absorbent walls were made. One is based on commercial micro-speakers, the other on a miniature MEMS transducer of similar dimensions but the conversion efficiency is an order of magnitude greater than conventional micro-speakers. Analytical modeling has shown two ways of improvements that have been undertaken, the first on the removal of short circuits present at the transducer, the second on optimizing the force factor for improving the conversion efficiency of electro-mechanics. The results sound absorption obtained from the MEMS transducers show that the solution has a real interest in the low frequency range where conventional solutions are not very effective
Dacosta, Fernandes Benoit. "Etude des propriétés électroniques et vibrationnelles de nano-objets métalliques et hybrides par spectroscopie femtoseconde." Thesis, Bordeaux, 2015. http://www.theses.fr/2015BORD0047/document.
Full textElectronic and vibrational dynamics have been studied in metallic and hybrid nano-object using femtosecond timeresolved spectroscopy. The study of electron-lattice energy exchanges in two-dimensional metallic systems showed an acceleration of the energy transfer between electrons and phonons due to confinement. This acceleration is governed by the smallest dimension of the nano-objects (2D-silver nano triangles ) when it becomes smaller than 10 nm. We also studied the vibrational dynamics of metallic nanoparticles, bimetallic and metal-dielectric. We investigated the longitudinal and radial modes of vibration of gold bipyramids which depend on their size, and studied their evolution under silver deposition. A high sensitivity of periods and amplitudes for small deposition were demonstrated. Our work on the evolution of acoustic vibrations of core-shell nano-objects (silver-silica and goldsilica) allowed us to obtain information on the quality of mechanical contact at the metal-dielectric interface. Finally, we studied the electronic interactions in hybrid metal / semiconductor (ZnO-Ag) nano-systems, and especially the charge transfer and energy exchanges between the two components. We showed a strong influence of the environment on the electron dynamics of ZnO and proved the existence of an electron transfer, photoinduced by an infrared pulse, from the metal particle to the semiconductor conduction band
Montoya, Maxime. "Sécurité adaptative et énergétiquement efficace dans l’Internet des Objets." Thesis, Lyon, 2019. http://www.theses.fr/2019LYSEM032.
Full textThe goal of this work is to propose new methods that provide both a high security and a high energy efficiency for integrated circuits for the IoT.On the one side, we study the security of a mechanism dedicated to energy management. Wake-up radios trigger the wake-up of integrated circuits upon receipt of specific wake-up tokens, but they are vulnerable to denial-of-sleep attacks, during which an attacker replays such a token indefinitely to wake-up a circuit and deplete its battery. We propose a new method to generate unpredictable wake-up tokens at each wake-up, which efficiently prevents these attacks at the cost of a negligible energy overhead.On the other side, we improve on the energy efficiency of hardware countermeasures against fault and side-channel attacks, with two different approaches. First, we present a new combined countermeasure, which increases by four times the power consumption compared to an unprotected implementation, introduces no performance overhead, and requires less than 8 bits of randomness. Therefore, it has a lower energy overhead than existing combined protections. It consists in an algorithm-level power balancing that inherently detects faults. Then, we propose an adaptive implementation of hardware countermeasures, which consists in applying or removing these countermeasures on demand, during the execution of the protected algorithm, in order to tune the security level and the energy consumption. A security evaluation of all the proposed countermeasures indicates that they provide an efficient protection against existing hardware attacks
Izquierdo, Christian. "Conception et réalisation d’un front-end analogique pour un récepteur multistandard multi-mode." Thesis, Lille 1, 2011. http://www.theses.fr/2011LIL10013/document.
Full textThe aim of this thesis is the design of a wide-band RF analogue front-end for a cellular multi-standard and multiband receiver. The main limitation of receivers today is the need of external SAW filters to protect the LNA and mixer from out-of-band interferers. In this thesis we propose a new technique of positive feed-back which transposes the filters transfer function in RF. Thus, a selective RF filter is created in the LNA input to improve non-linearity performances of the receiver. This RF filter is also configurable in bandwidth and center frequency. An experimental prototype has been made in 65nm CMOS technology. The positive feed-back improves the LNA out-of-band IIP3 by 17dB. The center frequency is configurable from 1.3GHZ to 2.85GHz. Power consumption is of 30mW, while maximal NF is of 6.5dB
Wong, Guillaume. "Développement de capteurs à ondes élastiques de surface auto-encapsulés aux fréquences intermédiaires pour environnement haute température." Thesis, Bourgogne Franche-Comté, 2018. http://www.theses.fr/2018UBFCD040.
Full textThis thesis contributes to scientific and technological research in the context of the ALCASAR and LANCASTER collaborative projects carried out in partnership with the companies Rakon, Frecnsys, Siltronix, SENSeOR and Claude Bernard University (Lyon 1). The challenge of ALCASAR is to produce surface elastic wavelength components (SAW) based on epitaxial AlN on sapphire for two types of applications. The first part will be to maximize the Qf product of resonators for frequency sources with noise of on the other hand, passive sensors interrogated at a distance for physical measurements at very high temperatures (above 650 ° C.). The frequency ranges of these components are located between 1 GHz and 4.6 GHz. LANCASTER 's stake is fundamentally similar, the differences being the material used for the substrate (Langatate) and the frequency ranges (500 MHz and 1 GHz). In the intended applications, WLP: Wafer Level packaging is provided and, in the specific case of the high temperature sensor. The main objective of the research is to optimize the response to the elastic wave components with respect to the differential stresses induced by the encapsulation when the environmental conditions, and in particular the temperature, vary
Issenmann, B. "Déformations d'interfaces fluides par la pression de radiation acoustique." Phd thesis, Université Sciences et Technologies - Bordeaux I, 2007. http://tel.archives-ouvertes.fr/tel-00283449.
Full textRenzi, Cédric. "Identification Expérimentale de Sources vibratoires par Résolution du problème Inverse modélisé par un opérateur Eléments Finis local." Phd thesis, INSA de Lyon, 2011. http://tel.archives-ouvertes.fr/tel-00715820.
Full textFraga, Da Silva Thiago. "Reducing development costs of large vocabulary speech recognition systems." Thesis, Paris 11, 2014. http://www.theses.fr/2014PA112232/document.
Full textOne of the outstanding challenges in large vocabulary automatic speech recognition (ASR) is the reduction of development costs required to build a new recognition system or adapt an existing one to a new task, language or dialect. The state-of-the-art ASR systems are based on the principles of the statistical learning paradigm, using information provided by two stochastic models, an acoustic (AM) and a language (LM) model. The standard methods used to estimate the parameters of such models are founded on two main assumptions : the training data sets are large enough, and the training data match well the target task. It is well-known that a great part of system development costs is due to the construction of corpora that fulfill these requirements. In particular, manually transcribing the audio data is the most expensive and time-consuming endeavor. For some applications, such as the recognition of low resourced languages or dialects, finding and collecting data is also a hard (and expensive) task. As a means to lower the cost required for ASR system development, this thesis proposes and studies methods that aim to alleviate the need for manually transcribing audio data for a given target task. Two axes of research are explored. First, unsupervised training methods are explored in order to build three of the main components of ASR systems : the acoustic model, the multi-layer perceptron (MLP) used to extract acoustic features and the language model. The unsupervised training methods aim to estimate the model parameters using a large amount of automatically (and inaccurately) transcribed audio data, obtained thanks to an existing recognition system. A novel method for unsupervised AM training that copes well with the automatic audio transcripts is proposed : the use of multiple recognition hypotheses (rather than the best one) leads to consistent gains in performance over the standard approach. Unsupervised MLP training is proposed as an alternative to build efficient acoustic models in a fully unsupervised way. Compared to cross-lingual MLPs trained in a supervised manner, the unsupervised MLP leads to competitive performance levels even if trained on only about half of the data amount. Unsupervised LM training approaches are proposed to estimate standard back-off n-gram and neural network language models. It is shown that unsupervised LM training leads to additive gains in performance on top of unsupervised AM training. Second, this thesis proposes the use of model interpolation as a rapid and flexible way to build task specific acoustic models. In reported experiments, models obtained via interpolation outperform the baseline pooled models and equivalent maximum a posteriori (MAP) adapted models. Interpolation proves to be especially useful for low resourced dialect ASR. When only a few (2 to 3 hours) or no acoustic data truly matching the target dialect are available for AM training, model interpolation leads to substantial performance gains compared to the standard training methods