No description available.

This thesis presents the results of the analysis of the bat auditory system and covers various aspect of bat behaviour of interest to the field of ultrasonic imaging. These small mammals use complex acoustic calls to navigate, identify and catch small fiying prey, with apparent performance surpassing any current man-made systems. Therefore, the study of such mammals can introduce new concepts for ultrasonic imaging. This thesis presents a review of the classification of bats by their acoustic emissions. It gives a detailed description of their acoustic calls and also provides information on the underlying processes involved in the bat auditory system. An analysis of the possible applications of the techniques used by the bats is done, in comparison with current ultrasonic technology. The work describes various approaches to reproduce bat behaviour by giving a detailed description of the concepts and the modelling techniques used. A biologically inspired algorithm, Bio Inspired Range finding Algorithm (BIRA), is developed to mimic the underlying processes involved in the bat auditory system, when forming a detailed acoustic image. This algorithm is inspired from the analysis of previous bat auditory models one developed by Saillant et al. in 1993 named Spectrogram Correlation And Transformation receiver (SCAT) and another one from Matsuo et al in 2004. To fulfill the purpose of this thesis, BIRA is designed as a new way of forming images for ultrasonic systems. For evaluating the performance of the BIRA in an environment similar to any engineering applications, an ultrasonic ranging system has also been designed. This system simulates the interactions between the source/receiver device, the medium of propagation and the target structure. It is used to quantify the axial resolution achieved by the BIRA. In a realistic environment, the BIRA appears to be limited in achieving the resolution of the bats. The results presented in this thesis question the way bats form an acoustic image and more specifically, the formation of a detailed range-profile image. The results support the assumption that bats may have a simpler detection method, either based on the detection of the first echo or on the detection of other acoustic effects created by a flying insect, such as the wing flutter.

L’objectif de cette thèse était de développer un capteur spectrophotométrique pour quantifier le benzène, le toluène et les xylènes (BTX) en temps réel aux valeurs limites d’exposition professionnelles, c’est-à-dire 20 ppmv pour le toluène, 50 ppmv pour les xylènes et 1 ppmv pour le benzène. L’étude a été menée avec plusieurs couches sensibles, un matériau silicique massif synthétisé par le procédé sol-gel ou des films minces (moins de 5 µm) déposés sur quartz par trempage-retrait dans une suspension de nanoparticules de silice mésoporeuse ou de précurseurs permettant d’obtenir des films fins continus de silice mésoporeuse. Nous avons pu démontrer l’efficacité de la détection des BTX en temps réel. L’utilisation des films mésoporeux a permis d’atteindre une répétabilité correcte du capteur (écart type <10%). Néanmoins, ni ces films de silice, ni leur fonctionnalisation par des groupements méthyle n’ont permis de résoudre le problème posé par l’humidité. Seule l'utilisation d'un sécheur que nous avons implémenté dans le dispositif de mesure a permis la détection du toluène en présence d'humidité ambiante. Dans cette configuration, les interférences de 40 ppmv de butanone, acétone et éthanol ont également été éliminées. Finalement la conception d’une cellule multiplaques a permis d’atteindre les limites de détection de 1 ppmv à 267 nm pour le toluène, 1 ppmv à 274 nm pour le p-xylène et 5 ppmv à 252 nm pour le benzène. Enfin, cette limite de détection peut être abaissée à 1 ppmv à 190 nm pour le benzène / The objective of this thesis was to develop a spectrophotometric sensor to quantify benzene, toluene and xylenes (BTX) in real time at the level of the occupational exposure limit values, ie 20 ppmv for toluene, 50 Ppmv for xylenes and 1 ppmv for benzene. The study was carried out with several sensitive layers, a solid silicic material synthesized by the sol-gel process or thin films (less than 5 μm) deposited on quartz by dip-coating a suspension of nanoparticles or a precursor solution leading to mesoporous silica thin films. We were able to demonstrate the effectiveness of BTX detection in real time. The use of continuous mesoporous films has enabled a correct repeatability of the sensor (standard deviation <10%). Nevertheless, neither these silica films nor their functionalization by methyl groups have solved the problem posed by moisture. Only the use of a dryer that we implemented in the measuring device allowed the detection of toluene in the presence of ambient humidity. In this configuration, interference of 40 ppmv of butanone, acetone and ethanol was also eliminated. Finally, the design of a multi-plate cell allowed to reach detection limits of 1 ppmv at 267 nm for toluene, 1 ppmv at 274 nm for p-xylene and 5 ppmv at 252 nm for benzene. Finally, this limit of detection can be lowered to 1 ppmv at 190 nm for benzene