Anomaly detection generally refers to algorithmic procedures aimed at identifying relatively rare events in data sets that differ substantially from the majority of the data set to which they belong. In the context of data series generated by sensors mounted on mobile devices for non-destructive inspection and monitoring, anomalies typically identify defects to be detected, therefore defining the main task of this class of devices. In this case, a useful way of operationally defining anomalies is to look at their information content with respect to the background data, which is typically noisy and therefore easily masking the relevant events if unfiltered. In this thesis, a Kullback-Leibler (KL) Divergence filter is proposed to detect signals with relatively high information content, namely anomalies, within data series. The data is generated by using the model of a broad class of proximity sensors that apply to devices commonly used in engineering practice. This includes, for example, sensory devices mounted on mobile robotic devices for the non-destructive inspection of hazardous or other environments that may not be accessible to humans for direct inspection. The raw sensory data generated by this class of sensors is often challenging to analyze due to the prevalence of noise over the signal content that reveals the presence of relevant features, as for example damage in gas pipelines. The proposed filter is built to detect the difference of information content between the data series collected by the sensor and a baseline data series, with the advantage of not requiring the design of a threshold. Moreover, differing from the traditional filters which need the prior knowledge or distribution assumptions about the data, this KL Divergence filter is model free and suitable for all kinds of raw sensory data. Of course, it is also compatible with classical signal distribution assumptions, such as Gaussian approximation, for instance. Also, the robustness and sensitivity of the KL Divergence filter are discussed under different scenarios with various signal to noise ratios of data generated by a simulator reproducing very realistic scenarios and based on models of real sensors provided by manufacturers or widely accepted in the literature.

<p>  </p> <p>Sensing systems have evolved significantly in recent years as a result of several advances in a number of sensor manufacturing approaches. The proximity measuring of approaching objects is a challenging, costly, and critical operation that permits the detection of any impediments without coming into touch with them and causing an unfavorable occurrence. However, developing a flexible proximity sensor capable of operating throughout a wide range of object motion continues to be a difficulty. The current work describes a polymer-based sensor that makes use of a nanostructure composite as the sensing element. The sensor will be used in healthcare and automotive applications in the near future. Composites comprising Thermoplastic Polyurethane (TPU) and Carbon Nanotubes (CNTs) are capable of sensing the presence of an external item at a great distance. The sensor model's performance was then enhanced further by microfabricating an integrated model with a certain shape. The design and production techniques for the TPU/CNTs proximity sensor are basic, and the sensor's performance demonstrates repeatability, as well as high electrical sensitivity and mechanical flexibility. The sensing process is based on the comparison of stored charges at the composite film sensor to the sensor's base voltage. The sensor operates reliably across a detection range of 2-20 cm. Tunneling and fringing effects are used to explain substantial capacitance shifts as sensing mechanisms. The structure's fringing capacitance effect has been thoroughly examined using ANSYS Maxwell (Ansoft) FEA simulation, as the measurements perfectly confirm the simulation's sensitivity trend. A novel mathematical model of fringe capacitance and subsequent tests demonstrate that the distance between an item and the sensor may be determined. Additionally, the model argues that the change in capacitance is significantly influenced by sensor resistivity, with the starting capacitance varying between 0.045pF and 0.024pF in the range 103-105 mm. This analytical model would enable the sensor's sensitivity to be optimized.</p> <p>Additionally, a new generation of durable elastomeric materials is commercially accessible for 3D printing, allowing the development of an entirely new class of materials for wearable and industrial applications. By using functional grading and adjusting to diverse users, the mechanical reaction of soft 3D-printed objects may now be modified for increased safety and comfort. Additionally, electronics may be included into these 3D printed lattice and wearable structures to offer input on the movement of objects associated with healthcare devices as well as automotive components. Thus, in order to investigate the influence of additive manufacturing on the sensitivity of TPU/CNT sensors, samples with equal thickness and size but varied orientations are printed and compared to hot-press samples. Among the many 3D printed patterns, the [0,0] direction has the highest sensitivity, and may be used as an optimum method for increased sensitivity. In contrast to the hot-press samples, the 3D-printed TPU/CNT film features a crystalline network, which may aid in the passage of surface charges and hence increase capacitance changes.</p> <p>To have a better understanding which feature, and parameter can give us the most sensitivity we need to do an optimization. This will be accomplished by collecting experimental and computational results and using them as a basis for establishing a computationally and experimentally supported Genetic Algorithm Assisted Machine Learning (GAML) framework combined with artificial neural network (ANN) to develop TPU/CNT nanocomposite flexible sensors in which material characterizations will be coupled to strain, tactile, electronic and proximity characteristics to probe intermolecular interactions between CNTs and polymers. The proposed framework provides enhanced predictive capabilities by managing multiple sets of data gathered from physical testing (material characterization and sensor testing) and multi-fidelity numerical models spanning all lengths scales. The GAML-ANN framework will allow the concurrent optimization of processing parameters and structural features of TPU/CNT nanocomposites, enabling fabrication of high-performance, lightweight flexible sensor systems.</p> <p>Our suggested nanocomposite sensor establishes a new mainstream platform for ultrasensitive object perception, demonstrating a viable prototype for wearable proximity sensors for motion analysis and the automobile sector.</p>

Pour la grande majorité des grands herbivores vivant en régions tempérées, l'hiver est la saison la plus limitante à cause des effets combinés du manque de ressources et de la sévérité des conditions climatiques. Les espèces qui ne développent pas d'adaptations morphologiques ou physiologiques particulières pour faire face à la sévérité de l'hiver doivent mettre en place des mouvements et des tactiques d'utilisation de l'espace appropriés (comme par exemple les migrations). Plus spécifiquement, ces patrons d'utilisation de l'espace peuvent émerger à différentes échelles spatio-temporelles pour permettre aux individus d'accéder aux ressources dont ils ont besoin et d'échapper aux conditions climatiques défavorables. Ainsi, ces patrons d'utilisation de l'espace déterminent le sort des individus et la dynamique des populations. Une compréhension détaillée de la relation entre les facteurs limitants durant l'hiver et la réponse des animaux en termes de mouvements est prépondérante pour préserver et gérer les populations d'ongulés sauvages avec succès, tout spécialement dans le contexte actuel de changements climatiques rapides qui induisent des modifications importantes dans le paysage et la distribution des ressources (par exemple changement de patrons de couverture neigeuse en hiver). La comparaison de tactiques de mouvements sous différents scénarios de conditions environnementales, par exemple au moyen d'analyses à vaste échelle de l'aire de distribution de l'espèce cible, représente une approche pertinente pour mieux comprendre comment les mouvements des animaux répondent aux changements dans le paysage. Le chevreuil (Capreoluscapreolus) est une espèce modèle particulièrement intéressante pour étudier ces questions parce que la distribution de cette espèce couvre la plupart des pays d'Europe, grâce à sa forte plasticité écologique / For the majority of large ungulates living in temperate regions winter is the limiting season, because of the combined effects of lack of resource availability and severity of climatic conditions. Those species that did not develop any particular morphological and physiological adaptation to cope with winter severity may adopt movement and space use tactics instead (like e.g. migration). Specifically, these space use patterns may emerge at different spatiotemporal scales to allow individuals to accede the resources they need and escape unfavourable conditions, thus determining both individual fate and population dynamics. A detailed understanding of the relationship between limiting factors in wintertime, and individual movement response, is remarkable to preserve and manage wildlife successfully, especially in the context of fast-occurring climate change that induces important alterations in landscape and resource distribution (e.g. changes of snow cover patterns). The comparison of movement tactics under different environmental scenarios, e.g. by means of large –scale analysis at the species distribution range, represents a valuable approach to work in that direction and to assess the effects of landscape alteration on individual movement. Roe deer (Capreolus capreolus) is an excellent model species to investigate these issues, because its distribution range covers most of Europe, thanks to its high ecological plasticity. For those populations that live in northern and mountain environments, winter is the limiting season because roe deer lack any morphological and physiological adaptations to cope with winter severity. In spite of the adoption of specific movement tactics such as partial migration from summer to winter ranges, roe deer may still face limiting conditions in some areas of the distribution range exposed to winter severity