Les légumineuses sont une source riche pour l’alimentation humaine comme celle du bétail mais elles sont aussi nécessaires à une agriculture durable. Cependant, les fractions majeures des protéines de réserve dans la graine sont pauvres en acides aminés soufrés et peuvent être accompagné de facteurs antinutritionnels, ce qui affecte leur valeur nutritive. Dans ce cadre, Medicago truncatula est une espèce modèle pour l’étude du développement de la graine des légumineuses, et en particulier concernant la phase d’accumulation des protéines de réserve. Vu la complexité des graines de légumineuses, une connaissance approfondie de leur morphogenèse ainsi que la caractérisation des mécanismes sous-jacents au développement de l’embryon et au remplissage de la graine sont essentielles. Une étude de mutagenèse a permis d’identifier le facteur de transcription DOF1147 (DNA-binding with One Finger) appartenant à la famille Zn-finger, qui s’exprime dans l’albumen pendant la transition entre les phases d’embryogenèse et de remplissage de la graine. Lors de mon travail de thèse, il a été possible de générer plusieurs constructions pour l’analyse de l’expression de DOF1147 ainsi que de la protéine DOF1147. Un protocole efficace pour la transformation génétique stable de M. truncatula a été établi et des études de localisation subcellulaire ont montré que DOF1147 est une protéine nucléaire. Un arbre phylogénétique a révélé différents groupes de facteurs de transcription DOF avec des domaines conservés dans leur séquence protéique. L’analyse du promoteur in silico chez plusieurs gènes-cible potentiels de DOF1147 a identifié les éléments cis-régulateurs de divers facteurs de transcription ainsi que des éléments répondant aux auxines (AuxREs), ce qui suggère un rôle possible de l’auxine pendant le développement de la graine. Une étude in vitro du développement de la graine avec divers régimes hormonaux, a montré l’effet positif de l’auxine sur la cinétique du développement de la graine, que ce soit en terme de gain de masse ou de taille, plus fort avec l’ANA que l’AIB. Grâce à une approche cytomique de ces graines en développement nous avons, en plus, démontré l’effet de l’auxine sur la mise en place de l’endoreduplication. En effet, celle-ci est l’empreinte cytogénétique de la transition entre les phases de division cellulaire et d’accumulation de substances de réserve lors du développement de la graine. Dans son ensemble, ce travail a démontré que l’auxine module la transition entre le cycle mitotique et les endocycles chez les graines en développement de M. truncatula en favorisant la continuité des divisions cellulaires tout en prolongeant simultanément l’endoreduplication. / Legumes are not only indispensible for sustainable agriculture but are also a rich source of protein in food and feed for humans and animals, respectively. However, major proteins stored in legume seeds are poor in sulfur-containing amino acids, and may be accompanied by anti-nutritional factors causing low protein digestibility problems. In this regard, Medicago truncatula serves as a model legume to study legume seed development especially the phase of seed storage protein accumulation. As developing legume seeds are complex structures, a thorough knowledge of the morphogenesis of the seed and the characterization of regulatory mechanisms underlying the embryo development and seed filling of legumes is essential. Mutant studies have identified a DOF1147 (DNA-binding with One Finger) transcription factor belonging to the Zn-Finger family which was expressed in the endosperm at the transition period between embryogenesis and seed filling phase. During my PhD work, a number of transgene constructs were successfully generated for expression analysis of DOF1147 gene as well as the DOF1147 protein. A successful transformation protocol was also established for stable genetic transformation of M. truncatula. Subcellular localization studies have demonstrated that DOF1147 is a nuclear protein. A phylogenetic tree revealed different groups of DOF transcription factors with conserved domains in their protein sequence. In silico promoter analysis of putative target genes of DOF1147 identified cis-regulatory elements of various transcription factors along with auxin responsive elements (AuxREs) suggesting a possible role of auxin during seed development. A study of in vitro seed development under different hormone regimes has demonstrated the positive effect of auxin on kinetics of seed development in terms of gain in seed fresh weight and size, with NAA having a stronger effect than IBA. Using the cytomic approach, we further demonstrated the effect of auxin on the onset of endoreduplication in such seeds, which is the cytogenetic imprint of the transition between the cell division phase and the accumulation of storage products phase during seed development. As a whole, this work highlighted that the auxin treatments modulate the transition between mitotic cycles and endocycles in M. truncatula developing seeds by favouring sustained cell divisions while simultaneously prolonging endoreduplication.

This thesis deals by the design of 1/4 car model for testing vehicle dampers, which can be used to simulate the real suspension of a vehicle wheel (up to a maximum car weight of 1,970 kg) and the so-called linear wheel suspension. A linear mathematical 1/4 car model with 2 DOF (Degrees Of Freedom) and data from literature search are used to design and derive the basic parameters of the device. The thesis contains a description of the linear mathematical model and its outputs (acceleration of the sprung mass and forces acting on the sprung mass), description of designed device, descriptions of created simulations (static, modal and harmonic analysis in ANSYS Workbench 2020 R2) and conceptual design of the modifications this device for another possible use for testing of bicycles.

This work addresses the deep-learning-based 6 Degree-of-Freedom (DoF) pose estimation utilizing 3D cameras on an iPhone 13 Pro. The task of pose estimation is to estimate the spatial rotation and translation of an object given its 2D or 3D images. During the pose estimation network training process, a common way to expand the training dataset is to generate synthetic images, which requires the 3D mesh of the target object. Although several famous datasets provide the 3D object files, it is still a problem when one wants to generate a customized real-world object. The typical 3D scanners are mainly designed for industrial usage and are usually expensive. We investigated in this project whether the 3D cameras on Apple devices can replace the industrial 3D scanners in the pose estimation pipeline and what might influence the results during scanning. During the data synthesis, we introduced a pose sampling method to equally sample on a sphere. Random transformation and background images from the SUN2012 dataset are applied, and the synthetic image is rendered through Blender. We picked five testing objects with different sizes and surfaces. Each object is scanned both by front TrueDepth camera and rear Light Detection and Ranging (LiDAR) camera with the ‘3d Scanner App’ on iOS. The network we used is based on PVNet, which uses a pixel-wise voting scheme to find 2D keypoints on RGB images and utilizes uncertainty-driven Perspective-n-Point (PnP) to compute the pose. We achieved both quantitative and qualitative results for each instance. i) TrueDepth camera outperforms Light Detection and Ranging (LiDAR) camera in most scenarios, ii) when an object has less reflective surface and high-contrast texture, the advantage of TrueDepth is more obvious. We also picked three baseline objects from Linemod dataset. Although the average accuracy is lower than the original paper, the performance of our baseline instances shows a similar trend to the original paper’s results. In conclusion, we proved that the 3D cameras on iPhone are capable of the pose estimation pipeline. / Detta arbete tar upp den djupinlärningsbaserade 6 Degree-of-Freedom (DoF) poseringsuppskattning med 3D-kameror på en iPhone 13 Pro. Uppgiften med poseuppskattning är att uppskatta den rumsliga rotationen och translationen av ett objekt givet dess 2D- eller 3D-bilder. Ett vanligt sätt att utöka träningsdataup- psättningen under träningsprocessen för positionsuppskattning är att generera syntetiska bilder, vilket kräver 3D-nätet för målobjektet. Även om flera kända datamängder tillhandahåller 3D-objektfilerna, är det fortfarande ett problem när man vill generera ett anpassat verkligt objekt. De typiska 3D-skannrarna är främst designade för industriell användning och är vanligtvis dyra. Vi undersökte i detta projekt om 3D-kamerorna på Apple-enheter kan ersätta de industriella 3D-skannrarna i poseskattningspipelinen och vad som kan påverka resultaten under skanning. Under datasyntesen introducerade vi en posesamplingsmetod för att sampla lika mycket på en sfär. Slumpmässig transformation och bakgrundsbilder från SUN2012-datauppsättningen tillämpas, och den syntetiska bilden renderas genom Blender. Vi valde ut fem testobjekt med olika storlekar och ytor. Varje objekt skannas både av den främre TrueDepth-kameran och den bakre ljusdetektions- och avståndskameran (LiDAR) med "3d-skannerappenpå iOS. Nätverket vi använde är baserat på PVNet, som använder ett pixelvis röstningsschema för att hitta 2D-nyckelpunkter på RGB-bilder och använder osäkerhetsdrivet Perspective-n-Point (PnP) för att beräkna poseringen. Vi uppnådde både kvantitativa och kvalitativa resultat för varje instans. i) TrueDepth-kameran överträffar Light Detection and Ranging-kameran (LiDAR) i de flesta scenarier, ii) när ett objekt har mindre reflekterande yta och högkontraststruktur är fördelen med TrueDepth högre. Vi valde också tre baslinjeobjekt från Linemod dataset. Även om den genomsnittliga noggrannheten är lägre än originalpapperet, visar prestandan för våra baslinjeinstanser en liknande trend som originalpapperets resultat. Sammanfattningsvis bevisade vi att 3D-kamerorna på iPhone är kapabla att göra positionsuppskattning.

No description available.

The world is slowly moving into increased human-robot interaction where both humans and robots can co-exist in the same domain. For the robot to be able to operate effectively in a man’s designed environment, it becomes necessary to model the robot with human capabilities as humans are seen as more capable. Replicating human becomes a huge challenge due to numerous degrees-of-freedom (DOFs) that human possess resulting into too many variables and nonlinear equations. Other challenges do occur like singularities.   In this thesis, the singularity challenge of a redundant humanoid arm is explored while maintaining a simple 7 DOF serial chain structure. As opposed to the 30 DOF human arm, a simpler 7 DOF humanoid arm is adopted and studied to eliminate the singularity challenges. The singularity problem mainly comes from the elbow and the spherical joints at the shoulder and wrist. A step-by-step review of available inverse kinematics techniques is made with more focus on the iterative Jacobian-based methods. A step-by-step approach is adopted so as to identify the source of singularities while using the iterative Jacobian-based techniques that are able to handle the nonlinearities of the equations.   The Singular Value Filtering (SVF) technique coupled with Selectively Damped Least Squares (SDLS) is employed. Without any restrictions to the stretch of the arm or end-effector pose, the method demonstrates, in conjunction with Euler angle singularity avoidance method, the elimination of singularity problems. This is achieved with no adjustment to kinematic model of the manipulator.

This thesis reports the development of a wide-bandwidth high-sensitivity mode-decoupled MEMS gyroscope showing robustness against ambient pressure variations. The designed gyroscope is based on a novel 2 degrees of freedom (DoF) sense mode oscillator, which allows increasing the operation bandwidth to the amount required by tactical-grade and inertial-grade operations while reaching the mechanical sensitivity of near matched-mode vibratory gyroscopes. Thorough theoretical study and finite element simulations verify the high performance operation of the proposed 2 DoF sense mode oscillator design. The designed gyroscope is fabricated using the in-house developed silicon-on-glass (SOG) micromachining technology at METU Microelectronics (METU-MET) facilities. The fabricated gyroscope measures only 5.1 x 4.6 mm square. The drive mode oscillator of the gyroscope reaches quality factor of 8760 under 25 mTorr vacuum environment, owing to high quality single crystal silicon structural layer. The sense mode bandwidth is measured to reach 2.5 kHz at 40 V proof mass voltage. When the fabricated gyroscope is operated with a relatively wide bandwidth of 1kHz, measurements show a relatively high raw mechanical sensitivity of 131 uV/(deg/s). Fabricated gyroscope is hybrid connected to external closed-loop drive mode amplitude control and open-loop sense mode readout electronics developed at METU-MEMS research group, to form a complete angular rate measurement system (ARMS). The scale factor of the ARMS is measured to be 13.1 mV/(deg/s) with a maximum R square nonlinearity of 0.0006 % and a maximum percent deviation nonlinearity of 0.141 %, while the maximum deviation of the scale factor for large vacuum level variations between 40 mTorr to 500 mTorr is measured to be only 0.38 %. The bias stability and angle random walk of the gyroscope are measured to be 131 deg/h and 1.15 deg/ rooth, respectively. It is concluded that, the mechanical structure can be optimized to show its theoretical limits of sensitivity with improvements in fabrication tolerances. The proposed 2 DoF sense mode oscillator design shows the potential of tactical-grade operation, while demonstrating extreme immunity to ambient pressure variations, by utilizing an optimized mechanical structure and connecting the gyroscope to dedicated low-noise electronics.

Growing needs for precise manipulation in medical surgery, manufacturing automation and structural health monitoring have motivated development of high accuracy, bandwidth and cost-effective sensing systems. Among these is a class of multi-axis electromagnetic devices where embedded magnetic fields can be capitalized for compact position estimation eliminating unwanted friction, stiction and inertia arising from dedicated and separate sensing mechanisms. Using fields for position measurements, however, is a challenging 'inverse problem' since they are often modeled in the 'forward' sense and their inverse solutions are often highly non-linear and non-unique. A general method to design a multisensor system that capitalizes on the existing magnetic field in permanent magnet (PM) actuators is presented. This method takes advantage of the structural field symmetry and meticulous placement of sensors to discretize the motion range of a PM-based device into smaller magnetic field segments, thereby reducing the required characterization domain. Within these localized segments, unique field-position correspondence is induced using field measurements from a network of multiple-axis sensors. A direct mapping approach utilizing trained artificial neural networks to attain multi-DOF positional information from distributed field measurements is employed as an alternative to existing computationally intensive model based methods which are unsuitable for real-time control implementation. Validation and evaluation of this technique are performed through field simulations and experimental investigation on an electromagnetic spherical actuator. An inclinometer was used as a performance comparison and experimental results have corroborated the superior tracking ability of the field-based sensing system. While the immediate application is field-based orientation determination of an electromagnetic actuator, it is expected that the design method can be extended to develop other sensing systems that harnesses other scalar, vector and tensor fields.

An algorithm for generating communicative, human-like motion for social humanoid robots was developed. Anticipation, exaggeration, and secondary motion were demonstrated as examples of communication. Spatiotemporal correspondence was presented as a metric for human-like motion, and the metric was used to both synthesize and evaluate motion. An algorithm for generating an infinite number of variants from a single exemplar was established to avoid repetitive motion. The algorithm was made task-aware by including the functionality of satisfying constraints. User studies were performed with the algorithm using human participants. Results showed that communicative, human-like motion can be harnessed to direct partner attention and communicate state information. Furthermore, communicative, human-like motion for social robots produced by the algorithm allows humans partners to feel more engaged in the interaction, recognize motion earlier, label intent sooner, and remember interaction details more accurately.

This thesis has been motivated by the growing needs for multi-degree of freedom (M-DOF) electromagnetic actuators capable of smooth and accurate multi-dimensional driving motions. Because high coercive rare-earth permanent-magnets (PMs) are widely available at low cost, their uses for developing compact, energy-efficient M-DOF actuators have been widely researched. To facilitate design analysis and optimization, this thesis research seeks to develop a general method based on distributed source models to characterize M-DOF PM-based actuators and optimize their designs to achieve high torque-to-weight performance with compact structures To achieve the above stated objective, a new method that is referred to here as distributed multi-level current (DMC) utilizes geometrically defined point sources has been developed to model electromagnetic components and phenomena, which include PMs, electromagnets (EMs), iron paths and induced eddy current. Unlike existing numerical methods (such as FEM, FDM, or MLM) which solve for the magnetic fields from Maxwell’s equations and boundary conditions, the DMC-based method develops closed-form solutions to the magnetic field and force problems on the basis of electromagnetic point currents in a multi-level structure while allowing trade-off between computational speed and accuracy. Since the multi-level currents can be directly defined at the geometrically decomposed volumes and surfaces of the components (such as electric conductors and magnetic materials) that make up of the electromagnetic system, the DMC model has been effectively incorporated in topology optimization to maximize the torque-to-weight ratio of an electromechanical actuator. To demonstrate the above advantages, the DMC optimization has been employed to optimize the several designs ranging from conventional single-axis actuators, 2-DOF linear-rotary motors to 3-DOF spherical motors. The DMC modeling method has been experimentally validated and compared against published data. While the DMC model offers an efficient means for the design analysis and optimization of electromechanical systems with improved computational accuracy and speed, it can be extended to a broad spectrum of emerging and creative applications involving electromagnetic systems.

The work presented in this thesis has two distinct parts: (i) development of a spinal motion measurement technique and (ii) incorporation of the spinal motion measurement with galvanic vestibular stimulation (GVS) technology, acting as a balance assist device hereafter referred to as a galvanic vestibular stimulation device (GVSD). The developed spinal motion measurement technique fulfills seven desired attributes: accuracy, portability, real-time data capture of dynamic data, non-invasive, small device footprint, clinically useful and of non-prohibitive cost. Applications of the proposed system range from diagnosis of spine injury to postural and balance monitoring, on-field as well as in the lab setting. The system is comprised of three inertial measurement sensors, respectively attached and calibrated to the head, torso and hips, based on the subject’s anatomical planes. Sensor output is transformed into meaningful clinical parameters of rotation, flexion-extension and lateral bending of each body segment with respect to a global reference space, then collected and visualized via an interactive graphical user interface (GUI). The accuracy of the proposed sensing system has been successfully verified with subject trials using a VICON optical motion measurement system. Next, the proposed motion measurement system and technique has been used to record a standing iv subject’s motion response to GVS. The data obtained allows the development of a new GVSD with the attributes of: eligibility for commercial licensing, portability, and capable of safely providing controlled stimulating current to the mastoid bones at varying levels and frequencies. The successful combination of the spinal motion measurement technique and GVSD represents the preliminary stage of a balance prosthesis.