No description available.

Vision-based grasp planning can be approached as an optimization problem, where a hand configuration that indicates a stable grasp needs to be located in a large search space. In this thesis, we proposed applying genetic algorithm (GA) to grasp planning of 3D object in arbitrary shapes and any robot hand. Details are given on the selection of operators and parameters of GA. GraspIt! simulator [2] is used for implementing the proposed algorithm and as the test environment. A quantitative analysis including the comparison with simple random algorithm and simulated annealing (SA) method is carried out to evaluate the performance of the GA based planner. Both GA and SA grasp planner are tested on different sets of hand-object. And two different quality metrics are used in the planning. Given the same amount of time, GA is shown to be capable of finding a force-closure grasp with higher stability than SA.

Les aérosols sont une composante très variable de l’atmosphère terrestre et font l’objet d’une attention croissante de la communauté scientifique et de la société. Depuis 2005, le LOA développe une activité reconnue en instrumentation, observation et inversion LiDAR pour la mesure des profils verticaux des paramètres descriptifs de ces aérosols. Depuis 2011, cette nouvelle thématique est soutenue par le projet européen ACTRIS (Aerosol Cloud and Trace gas Infrastructure) et le Labex CaPPA (Chemical And Physical Processes in the Atmosphere). Le premier objectif de cette thèse visait à déployer un nouveau LiDAR multi-longueur d’onde-Raman-polarisé, LILAS, d’une part sur la plateforme de mesures atmosphériques de l’université de Lille et d’autre part sur la station de géophysique de l’IRD à Dakar dans le cadre de la campagne de terrain SHADOW-2. Le second objectif visait à restituer puis étudier les propriétés optiques et microphysiques des couches aérosols détectées. Une méthodologie d’inversion innovante GARRLIC/GRASP a été mise en oeuvre et améliorée pour interpréter une série d’évenements aérosols (pollution locale, poussières minérales d’origine désertique transportées jusqu’à Lille mais également observés à Dakar, proche des zones sources). Cette nouvelle technique d’inversion combine les mesures primaires issues de la photométrie solaire (epaisseur optique et luminance spectrales) avec les profils de retrodiffusion LiDAR à 355, 532 et 1064 nm. Les propriétés des aérosols ambiants étant également fonction de l’humididité atmosphérique, une dernière partie a porté sur la mesure du profil de rapport de mélange de la vapeur d’eau accessible à partir de LILAS. / Aerosol particles are a highly variable component of the atmosphere and are now studied by a wide community. Since 2005, LOA is developing a recognized expertize in LiDAR observation devoted to aerosols profiling. Since 2011 this activity is supported by ACTRIS (Aerosol Cloud and Trace gas Infrastructure) and CaPPA (Chemical And Physical Processes in the Atmosphere) projects. The first objective of this work was to built, set up and characterize a new multi-wavelength Raman Polarized LiDAR (LILAS) operating at LOA observation platform located on the Campus. This system has also been operated during the SHADOW-2 field campaign (2015-2016) at Dakar (IRD station, Senegal). The second objective of the thesis consisted in developing aerosols retrievals and analyzing aerosols retrievals in term of optical and microphysical properties. An innovating synergetic approach (GARRLIC/GRASP) has been used and improved to interpret several aerosol events (local pollution, mineral dust transported to Lille and mineral dust detected at Dakar, close to sources). This new technique is combining primary data obtained from sun/sky photometer (spectral AOD and spectral radiance) and elastic LiDAR backscattering profiles (355, 532 and 1064 nm). Ambient aerosols properties being also sensitive to the atmospheric humidity, a last part of the work has be devoted to profiling water vapor mixing ratio from LILAS night time data.

Object recognition and manipulation are critical for enabling robots to operate within a household environment. There are many grasp planners that can estimate grasps based on object shape, but these approaches often perform poorly because they miss key information about non-visual object characteristics, such as weight distribution, fragility of materials, and usability characteristics. Object model databases can account for this information, but existing methods for constructing 3D object recognition databases are time and resource intensive, often requiring specialized equipment, and are therefore difficult to apply to robots in the field. We present an easy-to-use system for constructing object models for 3D object recognition and manipulation made possible by advances in web robotics. The database consists of point clouds generated using a novel iterative point cloud registration algorithm, which includes the encoding of manipulation data and usability characteristics. The system requires no additional equipment other than the robot itself, and non-expert users can demonstrate grasps through an intuitive web interface with virtually no training required. We validate the system with data collected from both a crowdsourcing user study and a set of grasps demonstrated by an expert user. We show that the crowdsourced grasps can produce successful autonomous grasps, and furthermore the demonstration approach outperforms purely vision-based grasp planning approaches for a wide variety of object classes.

Made available in DSpace on 2014-06-11T19:30:50Z (GMT). No. of bitstreams: 0 Previous issue date: 2012-05-04Bitstream added on 2014-06-13T18:40:58Z : No. of bitstreams: 1 utida_ma_dr_ilha.pdf: 1354826 bytes, checksum: 6130b79cafcdbe010b4b1482ede44474 (MD5) / Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) / O problema de carregamento de contêiner consiste em carregar um número conhecido de caixas de tipos diferentes e dimensões conhecidas dentro de um contêiner, de modo a utilizar o espaco do contêiner da melhor maneira possível, ou seja, de modo a diminuir os espacos não ocupados. O problema de carregamento de contêiner e um problema NP-difícil e, portanto, muito complexo de ser resolvido de forma exata usando um modelo matemático de programacao inteira e pacotes comerciais de otimizacao baseados em técnicas tipo branch and bound. Neste trabalho e realizada a elaboracao, assim como a implementacao computacional, de dois algoritmos de otimizacao. A primeira proposta de otimizacao consiste de um algoritmo heurístico construtivo. A segunda proposta de otimizacao um algoritmo GRASP que usa, na fase construtiva, uma generalizacao do algoritmo heurístico construtivo desenvolvido neste trabalho. A metaheurística GRASP desenvolvida neste trabalho apresentou resultados promissores quando foram realizados testes usando 13 instâncias conhecidas na literatura especializada / The container loading problem is to load a known number boxes of different dimensions and known within a container, in order to use the space of the container the best way possible, ie, in order to reduce the empty spaces. The problem of container loading is NP-hard and therefore too complex to be solved using an exact mathematical model integer programming and commercial packages based on optimization techniques branch and bound type. In this work the preparation, as well as the computational implementation of two optimization algorithms. The first proposal consists of an optimization constructive heuristic algorithm. The second proposal is an optimization GRASP that uses in construction phase, a generalization of the algorithm constructive heuristic developed in this work. The GRASP developed in this work showed promising results when tests were performed using 13 known instances in the literature specialized

Orientador: Rubén Augusto Romero Lázaro / Banca: José Roberto Sanches Mantovani / Banca: Silvio Alexandre de Araujo / Banca: José Elias Claudio Arroyo / Banca: Eduardo Nobuhiro Asada / Resumo: O problema de carregamento de contêiner consiste em carregar um número conhecido de caixas de tipos diferentes e dimensões conhecidas dentro de um contêiner, de modo a utilizar o espaco do contêiner da melhor maneira possível, ou seja, de modo a diminuir os espacos não ocupados. O problema de carregamento de contêiner e um problema NP-difícil e, portanto, muito complexo de ser resolvido de forma exata usando um modelo matemático de programacao inteira e pacotes comerciais de otimizacao baseados em técnicas tipo branch and bound. Neste trabalho e realizada a elaboracao, assim como a implementacao computacional, de dois algoritmos de otimizacao. A primeira proposta de otimizacao consiste de um algoritmo heurístico construtivo. A segunda proposta de otimizacao um algoritmo GRASP que usa, na fase construtiva, uma generalizacao do algoritmo heurístico construtivo desenvolvido neste trabalho. A metaheurística GRASP desenvolvida neste trabalho apresentou resultados promissores quando foram realizados testes usando 13 instâncias conhecidas na literatura especializada / Abstract: The container loading problem is to load a known number boxes of different dimensions and known within a container, in order to use the space of the container the best way possible, ie, in order to reduce the empty spaces. The problem of container loading is NP-hard and therefore too complex to be solved using an exact mathematical model integer programming and commercial packages based on optimization techniques branch and bound type. In this work the preparation, as well as the computational implementation of two optimization algorithms. The first proposal consists of an optimization constructive heuristic algorithm. The second proposal is an optimization GRASP that uses in construction phase, a generalization of the algorithm constructive heuristic developed in this work. The GRASP developed in this work showed promising results when tests were performed using 13 known instances in the literature specialized / Doutor

Made available in DSpace on 2018-08-02T00:03:44Z (GMT). No. of bitstreams: 1 tese_10270_Ata de defesa.pdf: 604618 bytes, checksum: 05ff69e59556087b1803468b45af37f1 (MD5) Previous issue date: 2016-09-22 / ...

The human hands are one of the most complex organs of the human body. As they enable us to grasp various objects in many different ways, they have played a crucial role in the rise of the human race. Being able to control hands as a human do is a key step towards friendly human-robots interaction and realistic virtual human simulations. Grasp generation has been mostly studied for the purpose of generating physically stable grasps. This paper addresses a different aspect: how to generate realistic, natural looking grasps that are similar to human grasps. To simplify the problem, the wrist position is assumed to be known and only the finger pose is generated. As the realism of a grasp is not easy to put into equations, data-driven machine learning techniques are used. This paper investigated the application of the deep neural networks to the grasp generation problems. Two different object shape representations (point cloud and multi-view depth images), and multiple network architectures are experimented, using a collected human grasping dataset in a virtual reality environment. The resulting generated grasps are highly realistic and human-like. Though there are sometimes some finger penetrations on the object surface, the general poses of the fingers around the grasped objects are similar to the collected human data. The good performance extends to the objects of categories previously unseen by the network. This work has validated the efficiency of a data-driven deep learning approach for human-like grasp synthesis. I believe the realistic-looking objective of the grasp synthesis investigated in this thesis can be combined with the existing mechanical, stable grasp criteria to achieve both natural-looking and reliable grasp generations. / Den mänskliga handen är en av de mest komplexa organen i människokroppen. Eftersom våra händer gör det möjligt för oss att hantera olika föremål på många olika sätt, har de spelat en avgörande roll i människans utveckling. Att kunna styra händer är ett viktigt steg mot interaktion mellan människor och robotar, samt skapa realistiska simuleringar av virtuella människor. Virtualla handgrepp har tidigare mest studerats för att generera fysiskt stabila grepp. I detta papper behandlas en annan aspekt: hur man genererar realistiska grepp som liknar en människas grepp. För att förenkla problemet antas att handledspositionen är känd och endast fingerkonfigurationen genereras. Eftersom hur realistiskt ett grepp är inte är lätt att beskriva i ekvationer, används istället data-driven maskininlärningsteknik. Detta papper undersöker tillämpningen av djupa neurala nätverken (Deep Neural Networks) för att generera grepp. Två olika representationer av former i 3D (punktmoln och bilder med djupinformation) och flera alternativa nätverksarkitekturer utvärderas med hjälp av en databas av mänskliga grepp samlad i en virtuell verklighetsmiljö. De resulterande genererade greppen är mycket realistiska och mänskliga. Även om det ibland förekommer något finger som penetrerar objektet, liknar den allmänna positioneringen av fingrarna den insamlade mänskliga datan. Denna goda prestanda gäller även för föremål i kategorier som aldrig tidigare setts av nätverket. I arbetet valideras också effektiviteten av ett data-drivet tillvägagångssätt baserat på djupa neurala nätverk för människoliknande syntes av grepp

No description available.

Telecommunication networks form an integral part of life. Avoiding failures on these networks is always not possible. Designing network structures that survive these failures have become important in ensuring the reliability of these network structures. With the introduction of SONET (Synchronous Optical Network) technology, rings have become the preferred survivable network structure. This network configuration has a set of disjoint rings (each node being a part of single ring), and these disjoint rings are connected via another main ring. In this research, we present a mathematical model for the design of such disjoint rings with node number balance criterion among the rings. When, given a set of nodes and distances between them, the Balanced Disjoint Rings (BDR) problem is the minimum total link length clustering of nodes into a given number of disjoint rings in such a way that there is almost the same number of nodes in each ring. The BDR problem is a class of the standard Traveling Salesman Problem (TSP). It is clear from this observation that the BDR problem becomes a TSP when the number of rings required is set to one. Hence BDR is NP-Hard, and we do not expect to obtain a polynomial time algorithm for its solution. To overcome this problem, we developed a set of construction heuristics (Break-MST, Distance Method, Hybrid Method, GRASP-Based Distance Method) and improvement heuristics (Multi-Exchange, Single Move). Different combinations of construction and improvement heuristics were implemented and the quality of solution thus obtained was compared to the standard Branch and Cut Technique. It was found that the algorithm with GRASP-Based Distance Method as the construction heuristic and multi-exchange - single-move combination as the improvement heuristic performed better than other combinations. All combinations performed better in general than the standard Branch and Cut technique in terms of solution time.