Global ETD Search

261	EFFICIENTNEXT: EFFICIENTNET FOR EMBEDDED SYSTEMS Abhishek Rajendra Deokar (12456477) 12 July 2022 (has links) <p>Convolutional Neural Networks have come a long way since AlexNet. Each year the limits of the state of the art are being pushed to new levels. EfficientNet pushed the performance metrics to a new high and EfficientNetV2 even more so. Even so, architectures for mobile applications can benefit from improved accuracy and reduced model footprint. The classic Inverted Residual block has been the foundation upon which most mobile networks seek to improve. EfficientNet architecture is built using the same Inverted Residual block. In this paper we experiment with Harmonious Bottlenecks in place of the Inverted Residuals to observe a reduction in the number of parameters and improvement in accuracy. The designed network is then deployed on the NXP i.MX 8M Mini board for Image classification.</p> Digital processor architectures Convolutional Neural Network Coordinate Attention Adaptive Sharpness Aware Minimization Harmonious Bottleneck Sandglass Bottleneck Inverted Residual Fused-MBConv CIFAR-10 NXP i.MX 8M Mini GELUs activation function Computer Engineering
262	Análisis y mejora del comportamiento dimensional de termoplásticos impresos en 3D mediante modelado por deposición fundida sometidos a un proceso de tratamiento térmico Lluch Cerezo, Joaquín 03 July 2023 (has links) Tesis por compendio / [ES] En la actualidad, el modelado por deposición fundida (FDM) es el tipo de tecnología de fabricación aditiva más difundida y estudiada dada su facilidad de uso y economía de proceso. No obstante, debido a las anisotropías generadas durante el proceso aditivo, las piezas fabricadas mediante FDM presentan limitaciones en su uso funcional. Estas anisotropías dependen de los parámetros del proceso y pueden provocar variaciones dimensionales y cambios en las propiedades mecánicas de las piezas. Para mejorar dichas características, se puede recurrir a diversos post-procesos como el tratamiento térmico. Sin embargo, durante su aplicación se pueden producir variaciones dimensionales en las piezas tratadas que reduzcan o anulen su aplicabilidad industrial. En la presente Tesis Doctoral se ha abordado el estudio del comportamiento dimensional de piezas fabricadas mediante FDM sometidas a tratamiento térmico. Se ha evaluado una propuesta de mejora en el post-proceso consistente en el uso de un molde de polvo cerámico compactado alrededor de las piezas a fin de minimizar la aparición de deformaciones durante el tratamiento térmico. El estudio de las deformaciones durante el post-procesado térmico se ha focalizado en los dos materiales termoplásticos más empleados en FDM, el ácido poliláctico (PLA) y el acrilonitrilo butadieno estireno (ABS), de naturaleza semicristalina y amorfa respectivamente. Se han analizado las variaciones dimensionales sufridas por las piezas durante el tratamiento térmico, considerando la influencia de la orientación de las líneas depositadas, la temperatura del tratamiento y el uso del molde de polvo cerámico. Para evaluar la mejora aplicada en el post-procesado térmico, se ha definido y analizado la eficacia del molde con las mismas variables del estudio dimensional. A fin de poder predecir las deformaciones sufridas por las piezas tratadas y la eficacia del molde en un amplio rango de temperaturas, se ha realizado una aproximación polinómica con los resultados obtenidos experimentalmente. / [CA] Actualment, el modelatge per deposició fosa (FDM) és el tipus de tecnologia de fabricació additiva més difosa i estudiada atesa la facilitat d'ús i economia de procés. Això no obstant, a causa de les anisotropies generades durant el procés additiu, les peces fabricades mitjançant FDM presenten limitacions en el seu ús funcional. Aquestes anisotropies depenen dels paràmetres del procés i poden provocar variacions dimensionals i canvis a les propietats mecàniques de les peces. Per millorar aquestes característiques, es pot recórrer a diversos postprocessos com el tractament tèrmic. No obstant això, durant la seva aplicació es poden produir variacions dimensionals a les peces tractades que redueixin o anul·lin la seva aplicabilitat industrial. En aquesta Tesi Doctoral s'ha abordat l'estudi del comportament dimensional de peces fabricades mitjançant FDM sotmeses a tractament tèrmic. S'ha avaluat una proposta de millora en el post-procés consistent en l'ús d'un motlle de pols ceràmic compactat al voltant de les peces per tal de minimitzar l'aparició de deformacions durant el tractament tèrmic. L'estudi de les deformacions durant el post-processat tèrmic s'ha focalitzat en els dos materials termoplásticos més empleats en FDM, l'àcid poliláctico (PLA) i l'acrilonitrilo butadien estireno (ABS), de naturalesa semicristalina i amorfa respectivament. S'han analitzat les variacions dimensionals sofrides per les peces durant el tractament tèrmic, tenint en compte la influència de l'orientació de les línies dipositades, la temperatura del tractament i l'ús del motlle de pols ceràmic. Per avaluar la millora aplicada en el post-processat tèrmic, s'ha definit i analitzat l'eficàcia del motlle amb les mateixes variables de l'estudi dimensional. Per tal de poder predir les deformacions patides per les peces tractades i l'eficàcia del motlle en un ampli rang de temperatures, s'ha fet una aproximació polinòmica amb els resultats obtinguts experimentalment. / [EN] Nowadays, fused deposition modeling (FDM) is the most widespread and studied additive manufacturing technology due to its ease of use and process economics. However, due to the anisotropies generated during the additive process, FDM fabricated parts have limitations in their functional use. These anisotropies depend on the process parameters and can lead to dimensional variations and changes in the mechanical properties of the parts. Various post-processes, such as heat treatment, can be used to improve these characteristics. However, during its application, dimensional variations can occur in the treated parts that reduce or make their industrial applicability impossible. In this Thesis, the dimensional behavior of FDM parts subjected to heat treatment has been studied. A post-processing improvement proposal consisting of using a compacted ceramic powder mould around the parts to minimize deformations during the heat treatment has been evaluated. The study of deformations during thermal post-processing has focused on the two most widely used thermoplastic materials in FDM, polylactic acid (PLA) and acrylonitrile butadiene styrene (ABS), with semi-crystalline and amorphous nature respectively. For this purpose, standard specimens coded according to different internal geometries were manufactured. The dimensional variations suffered by the parts during the heat treatment have been analyzed considering the influence of the orientation of the deposited lines, the material used, the treatment temperature and the use of the ceramic powder mould. To evaluate the thermal post-processing improvement, the effectiveness of mould has been defined and analyzed with the same variables of the dimensional study. To predict the deformations suffered by the treated parts and the efficacy of the mould in a wide range of temperatures, a polynomial approximation has been fitted to the results obtained experimentally. / Lluch Cerezo, J. (2023). Análisis y mejora del comportamiento dimensional de termoplásticos impresos en 3D mediante modelado por deposición fundida sometidos a un proceso de tratamiento térmico [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/194607 / Compendio Modelado por deposición fundida (FDM) Tratamiento térmico postproceso Recocido Deformaciones térmicas de piezas Molde de polvo cerámico Acrilonitrilo butadieno estireno (ABS) Ácido poliláctico (PLA) Fused deposition modelling (FDM) Post-process thermal treatment Annealing Thermal part deformations Ceramic powder mould Acrylonitrile butadiene styrene (ABS) Polylactic acid (PLA)
263	Model metropolitní optické sítě / Model of the metropolitan optical network Prudík, Jiří January 2008 (has links) The purpose of this master’s thesis is foremost to provide a simple guide how to build elements of optical metropolitan area network. The basic model consists to sequence of construction, network topology, passive and active parts. The collection contains examples of alternative technology such as Wireless LAN with different frequency. The optical network construction based on optical cable, fibres, splices, trays, adapters, connectors and active parts for example a lot of media convertor models. After that there are demonstrating type of wavelength division multiplexer used in metropolitan area network – passive planar PCL splitter. One of the passive planar splitter are used to increase optical fibre channel. At the end of the collection a simplified examples of used measurements – optical time domain reflectometry and optical fibre transmission. Contains standard protocols or reflectogram. The conclusion of this thesis summarizes costs of FTTb (Fibre To The Building) model of optical metropolitan area network in Czech republic and future contribution for society.
264	Evaluation of Target Tracking Using Multiple Sensors and Non-Causal Algorithms Vestin, Albin, Strandberg, Gustav January 2019 (has links) Today, the main research field for the automotive industry is to find solutions for active safety. In order to perceive the surrounding environment, tracking nearby traffic objects plays an important role. Validation of the tracking performance is often done in staged traffic scenarios, where additional sensors, mounted on the vehicles, are used to obtain their true positions and velocities. The difficulty of evaluating the tracking performance complicates its development. An alternative approach studied in this thesis, is to record sequences and use non-causal algorithms, such as smoothing, instead of filtering to estimate the true target states. With this method, validation data for online, causal, target tracking algorithms can be obtained for all traffic scenarios without the need of extra sensors. We investigate how non-causal algorithms affects the target tracking performance using multiple sensors and dynamic models of different complexity. This is done to evaluate real-time methods against estimates obtained from non-causal filtering. Two different measurement units, a monocular camera and a LIDAR sensor, and two dynamic models are evaluated and compared using both causal and non-causal methods. The system is tested in two single object scenarios where ground truth is available and in three multi object scenarios without ground truth. Results from the two single object scenarios shows that tracking using only a monocular camera performs poorly since it is unable to measure the distance to objects. Here, a complementary LIDAR sensor improves the tracking performance significantly. The dynamic models are shown to have a small impact on the tracking performance, while the non-causal application gives a distinct improvement when tracking objects at large distances. Since the sequence can be reversed, the non-causal estimates are propagated from more certain states when the target is closer to the ego vehicle. For multiple object tracking, we find that correct associations between measurements and tracks are crucial for improving the tracking performance with non-causal algorithms. evaluation target tracking multiple sensors non-causal smoother smoothing tracking vehicle tracking camera lidar estimate estimation prediction vehicle dynamics sensor fusion real-time tracking extended kalman filter filter validation validation position estimation velocity estimation dynamic model model complexity multi object tracking multiple object tracking single object tracking data association tracking fundamentals iterated kalman filter track management gnn global nearest neighbour mahalanobis mahalanobis distance performance evaluation differential gps dgps roi ego several sensors sensors rmse root mean square error invertible motion anti-causal motion anti-causal tracking constant velocity gnn imu tfs two filter smoother ekf rts radar inertial measurement unit nonlinear nonlinear systems mono camera monocular camera noise model tracking performance fixed interval smoothing m/n logic centralized fusion non-causal object tracker car tracking car dynamics automotive active safety object tracking automotive industry thesis master reverse dynamics reverse tracking reverse sequence sequence tracking data propagation ground truth estimating ground truth additional sensors mounted sensors true estimates environment comparison algorithm independent targets overlapping measurements occluded track switch improve lower uncertainty more certain state process noise covariance sampling image sprt adas cnn cv pdf track target ego tracker tentative track observatiom online tracking offline tracking online offline recorded sequences robust self driving self-driving car traffic trajectory true state scenario scenarios future accurate output advanced driver assistance systems non-linear complex noise pedestrian truck bus maneuvering vehicles processed measurement frame state correction probability density function tuning likelihood transition measurement motion model recursion gaussian approximation distribution linear jacobian multiplicative noise ratio ad hoc ad hoc state space approach backward auction euclidean distance statistical threshold gating association margin normalize covariance matrix fusion confirmed rejected tentative history absolute error modular ego motion parameters variables logg hardware specification fused causal factorization independent uncorrelated transform moving rotation translation oncoming overtaking Control Engineering Reglerteknik

Page generated in 0.0351 seconds