Coordenação de robôs móveis é um tópico importante de pesquisa dado que existem tarefas que podem ser desenvolvidas de forma mais eficiente e com menor custo por um grupo de robôs do que por um só robô. Nesta dissertação é apresentado um estudo sobre coordenação de robôs móveis para o problema de navegação em ambientes externos. Para isso, foi desenvolvido um sistema de localização utilizando os dados de odometria e do receptor GPS, e um sistema de desvio de obstáculos para planejar a trajetória livre de obstáculos. Os movimentos coordenados foram realizados em função de um líder e qualquer robô da formação pode assumir a liderança. A liderança é assumida pelo robô que ultrapassar a distância mínima a um obstáculo. Movimentos estáveis são gerados através de uma lei de controle descentralizada baseada nas coordenadas dos robôs. Para garantir a estabilidade da formação quando há alternância de líder ou remoção de robôs, foi feito controle tolerante a falhas para um grupo de robôs móveis. O controle tolerante a falhas é baseado em controle H \'INFINITO\' por realimentação da saída de sistemas lineares sujeitos a saltos Markovianos para garantir a estabilidade da formação quando um dos robôs é perdido durante o movimento coordenado. Os resultados do sistema de localização mostram que o uso de filtro robusto para a fusão de dados produz uma melhor estimativa da posição do robô móvel. Os resultados também mostram que o sistema de desvio de obstáculos é capaz de gerar uma trajetória livre de obstáculos em ambientes desconhecidos. E por fim, os resultados do sistema de coordenação mostram que o grupo de robôs mantém a formação desejada percorrendo a trajetória de referência na presença de distúrbios ou quando um robô sai da formação. / Coordination of mobile robots is an important topic of research because there are tasks that may be too difficult for a single robot to perform alone, these tasks can be performed more efficiently and cheaply by a group of mobile robots. This dissertation presents a study on the coordination of mobile robots to the problem of navigation in outdoor environments. To solve this problem, a localization system using data from odometry and GPS receiver, and an obstacle avoidance system to plan the collision-free trajectory, were developed. The coordinated motions are performed by the robots that follow a leader, and any robot of the formation can assume the leadership. The leadership is assumed by a robot when it exceeds the threshold distance to an obstacle. Stable motions are generated by a decentralized control law based on the robots coordinates. To ensure the stability formation when there is alternation of leader or one of the robots is removed, we made a fault tolerant control for a group of mobile robots. The fault tolerant approach is based on output feedback H \'INFINITE\' control of Markovian jump linear systems to ensure stability of the formation when one of the robots is lost during the coordinated motion. The results of the localization system show that the use of robust filter for data fusion produces a better estimation of the mobile robots position. The results also show that the obstacle avoidance system is capable of generating a path free from obstacles in unknown environments. Finally, the results of the coordination system show that the group of robots maintain the desired formation along the reference trajectory in the presence of disturbance or removal of one of them.

Made available in DSpace on 2015-04-11T14:02:57Z (GMT). No. of bitstreams: 1 Afonso.pdf: 820074 bytes, checksum: 796ca447ff3c69734519173f92044438 (MD5) Previous issue date: 2013-03-27 / Wireless Sensor Networks (WSNs) can be used in environmental conservation applications and studies due to its wireless communication, sensing, and monitoring capabilities. In the Ecology context, amphibians are used as bioindicators of ecosystemic changes of a region and can early indicate environmental problems. Thus, biologists monitor the anuran (frogs and toads) population in order to establish environmental conservational strategies. Anuran were chosen because the sounds they emit allow classification by using microphones and signal processing. In this work we propose and evaluate some distributed algorithms for anuran classification based on their calls (vocalizations) in the habit using WSNs. This method is interesting because it is not intrusive and it allows remote monitoring. Our solution builds cluster of nodes whose acoustic collected measurements are correlated. The nodes of the same group are combined to generate local classification decisions. Then, these decisions are combined to generate a global decision. We use k-means algorithm for clustering nodes with correlated measurements, which groups instances by similarity. Experiments show that, in comparison with other literature algorithms, the error rate of our solution were 26 pp (percentage points) lower. / As Redes de Sensores Sem Fios (RSSFs) podem ser utilizadas em aplicações de conservação e estudo ambiental devido à sua capacidade de sensoriamento, monitoramento e comunicação sem fio. Dentro do contexto da Ecologia, os anfíbios são utilizados como bioindicadores de mudanças no ecossistema de uma região e podem precocemente indicar problemas ambientais. Desta forma, os biólogos monitoram a população de anuros (sapos e rãs) a fim de estabelecer estratégias de conservação do meio ambiente. Os anuros são escolhidos por causa sons que emitem (coaxar), que permitem a identificação dessas espécies por meio de microfones e processamento do sinal. Portanto, neste trabalho propomos e avaliamos alguns algoritmos distribuídos para classificação de anuros baseados em suas vocalizações em seu habitat usando RSSF. Este método é interessante pois não é intrusivo e permite o monitoramento remoto. Nossa solução cria grupos de nós sensores cujas medidas acústicas coletadas estão correlacionadas. Os dados dos nós de um mesmo grupo são combinados para gerar decisões de classificação locais. Essas decisões são então combinadas para formar uma decisão global. Para agrupar os nós com medidas correlacionadas, utilizamos o algoritmo k-means, que agrupa instâncias similares. Os experimentos mostram que, em comparação com outros algoritmos da literatura, a taxa de erro da nossa solução chegou ser até 26 pp (pontos percentuais) menor.

Orientador: Eurípedes Guilherme de Oliveira Nóbrega / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Mecânica / Made available in DSpace on 2018-08-23T01:49:32Z (GMT). No. of bitstreams: 1 Souza_PabloRodrigode_D.pdf: 10455045 bytes, checksum: b8461069008fc0a12ec07525f9bd8d2e (MD5) Previous issue date: 2013 / Resumo: O resumo poderá ser visualizado no texto completo da tese digital / Abstract: The abstract is available with the full electronic document / Doutorado / Mecanica dos Sólidos e Projeto Mecanico / Doutor em Engenharia Mecânica

The rapid increasing interest in wireless communication has led to the continuous development of wireless devices and technologies. The modern convergence and interoperability of wireless technologies has further increased the amount of services that can be provided, leading to the substantial demand for access to the radio frequency spectrum in an efficient manner. Cognitive radio (CR) an innovative concept of reusing licensed spectrum in an opportunistic manner promises to overcome the evident spectrum underutilization caused by the inflexible spectrum allocation. Spectrum sensing in an unswerving and proficient manner is essential to CR. Cooperation amongst spectrum sensing devices are vital when CR systems are experiencing deep shadowing and in a fading environment. In this thesis, cooperative spectrum sensing (CSS) schemes have been designed to optimize detection performance in an efficient and implementable manner taking into consideration: diversity performance, detection accuracy, low complexity, and reporting channel bandwidth reduction. The thesis first investigates state of the art spectrums sensing algorithms in CR. Comparative analysis and simulation results highlights the different pros, cons and performance criteria of a practical CSS scheme leading to the problem formulation of the thesis. Motivated by the problem of diversity performance in a CR network, the thesis then focuses on designing a novel relay based CSS architecture for CR. A major cooperative transmission protocol with low complexity and overhead - Amplify and Forward (AF) cooperative protocol and an improved double energy detection scheme in a single relay and multiple cognitive relay networks are designed. Simulation results demonstrated that the developed algorithm is capable of reducing the error of missed detection and improving detection probability of a primary user (PU). To improve spectrum sensing reliability while increasing agility, a CSS scheme based on evidence theory is next considered in this thesis. This focuses on a data fusion combination rule. The combination of conflicting evidences from secondary users (SUs) with the classical Dempster Shafter (DS) theory rule may produce counter-intuitive results when combining SUs sensing data leading to poor CSS performance. In order to overcome and minimise the effect of the counter-intuitive results, and to enhance performance of the CSS system, a novel state of the art evidence based decision fusion scheme is developed. The proposed approach is based on the credibility of evidence and a dissociability degree measure of the SUs sensing data evidence. Simulation results illustrate the proposed scheme improves detection performance and reduces error probability when compared to other related evidence based schemes under robust practcial scenarios. Finally, motivated by the need for a low complexity and minmum bandwidth reporting channels which can be significant in high data rate applications, novel CSS quantization schemes are proposed. Quantization methods are considered for a maximum likelihood estimation (MLE) and an evidence based CSS scheme. For the MLE based CSS, a novel uniform and optimal output entropy quantization scheme is proposed to provide fewer overhead complexities and improved throughput. While for the Evidence based CSS scheme, a scheme that quantizes the basic probability Assignment (BPA) data at each SU before being sent to the FC is designed. The proposed scheme takes into consideration the characteristics of the hypothesis distribution under diverse signal-to-noise ratio (SNR) of the PU signal based on the optimal output entropy. Simulation results demonstrate that the proposed quantization CSS scheme improves sensing performance with minimum number of quantized bits when compared to other related approaches.

Cette thèse aborde le problèeme du suivi et de la classification de plusieurs objets simultanément.Il est montré dans la thèese que les fonctions de croyance permettent d'améliorer les résultatsfournis par des méthodes classiques à base d'approches Bayésiennes. En particulier, une précédenteapproche développée dans le cas d'un seul objet est étendue au cas de plusieurs objets. Il est montréque dans toutes les approches multi-objets, la phase d'association entre observations et objetsconnus est fondamentale. Cette thèse propose également de nouvelles méthodes d'associationcrédales qui apparaissent plus robustes que celles trouvées dans la littérature. Enfin, est abordée laquestion de la classification multi-capteurs qui nécessite une seconde phase d'association. Dans cedernier cas, deux architectures de fusion des données capteurs sont proposées, une dite centraliséeet une autre dite distribuée. De nombreuses comparaisons illustrent l'intérêt de ces travaux, queles classes des objets soient constantes ou variantes dans le temps. / This thesis deals with multi-objet tracking and classification problem. It was shown that belieffunctions allow the results of classical Bayesian methods to be improved. In particular, a recentapproach dedicated to a single object classification which is extended to multi-object framework. Itwas shown that detected observations to known objects assignment is a fundamental issue in multiobjecttracking and classification solutions. New assignment solutions based on belief functionsare proposed in this thesis, they are shown to be more robust than the other credal solutions fromrecent literature. Finally, the issue of multi-sensor classification that requires a second phase ofassignment is addressed. In the latter case, two different multi-sensor architectures are proposed, aso-called centralized one and another said distributed. Many comparisons illustrate the importanceof this work, in both situations of constant and changing objects classes.

Autonomous reconnaissance missions are called for in extreme environments, as well as in potentially hazardous (e.g., the theatre, disaster-stricken areas, etc.) or inaccessible operational areas (e.g., planetary surfaces, space). Such future missions will require increasing degrees of operational autonomy, especially when following up on transient events. Operational autonomy encompasses: (1) Automatic characterization of operational areas from different vantages (i.e., spaceborne, airborne, surface, subsurface); (2) automatic sensor deployment and data gathering; (3) automatic feature extraction including anomaly detection and region-of-interest identification; (4) automatic target prediction and prioritization; (5) and subsequent automatic (re-) deployment and navigation of robotic agents. This paper reports on progress towards several aspects of autonomous (CISR)-I-4 systems, including: Caltech-patented and NASA award-winning multi-tiered mission paradigm, robotic platform development (air, ground, water-based), robotic behavior motifs as the building blocks for autonomous telecommanding, and autonomous decision making based on a Caltech-patented framework comprising sensor-data-fusion (feature-vectors), anomaly detection (clustering and principal component analysis), and target prioritization (hypothetical probing).

On étudie dans cette thèse des méthodes de localisation coopérative de robots mobiles sans utilisation de mesures extéroceptives relatives, comme des angles ou des distances entre robots. Les systèmes de localisation considérés sont basés sur des mesures de radionavigation sur des balises fixes ou des satellites. Pour ces systèmes, on observe en général un écart entre la position observée et la position réelle. Cet écart systématique (appelé biais) peut être dû à une mauvaise position de la balise ou à une différence entre la propagation réelles des ondes électromagnétiques par rapport aux conditions standard utilisées pour établir les modèles d’observation. L’influence de ce biais sur la localisation des robots est non négligeable. La coopération et l’échange de données entre les robots (estimations des biais, estimations des positions et données proprioceptives) est une approche qui permet de corriger ces erreurs systématiques. La localisation coopérative par échange des estimations est sujette aux problèmes de consanguinité des données qui peuvent engendrer des résultats erronés, en particulier trop confiants. Lorsque les estimations sont utilisées pour la navigation autonome à l’approche, on doit éviter tout risque de collision qui peut mettre en jeu la sécurité des robots et des personnes aux alentours. On doit donc avoir recours à un mécanisme d’intégrité vérifiant que l’erreur commise reste inférieure à une erreur maximale tolérable pour la mission. Dans un tel contexte, il est nécessaire de caractériser des domaines de confiance fiables contenant les positions des robots mobiles avec une forte probabilité. L’utilisation des méthodes ensemblistes à erreurs bornées est considérée alors comme une solution efficace. En effet, ce type d’approche résout naturellement le problème de consanguinité des données et fournit des domaines de confiance fiables. De surcroît, l’utilisation de modèles non-linéaires ne pose aucun problème de linéarisation. Après avoir modélisé un système coopératif de nr robots avec des mesures biaisées sur des balises, une étude d’observabilité est conduite. Deux cas sont considérés selon la nature des mesures brutes des observations. En outre, des conditions d’observabilité sont démontrées. Un algorithme ensembliste de localisation coopérative est ensuite présenté. Les méthodes considérées sont basées sur la propagation de contraintes sur des intervalles et l’inversion ensembliste. La coopération est effectuée grâce au partage des positions estimées, des biais estimés et des mesures proprioceptives.L’échange des estimations de biais permet de réduire les incertitudes sur les positions des robots. Dans un cadre d’étude simple, la faisabilité de l’algorithme est évaluée grâce à des simulations de mesures de distances sur balises en utilisant plusieurs robots. La coopération est comparée aux méthodes non coopératives. L’algorithme coopératif ensembliste est ensuite testé sur des données réelles en utilisant deux véhicules. Les performances de la méthode ensembliste coopérative sont enfin comparées avec deux méthodes Bayésiennes séquentielles, notamment une avec fusion par intersection de covariance. La comparaison est conduite en termes d’exactitude et d’incertitude. / In this work, we study some cooperative localization issues for mobile robotic systems that interact with each other without using relative measurements (e.g. bearing and relative distances). The considered localization technologies are based on beacons or satellites that provide radio-navigation measurements. Such systems often lead to offsets between real and observed positions. These systematic offsets (i.e, biases) are often due to inaccurate beacon positions, or differences between the real electromagnetic waves propagation and the observation models. The impact of these biases on robots localization should not be neglected. Cooperation and data exchange (estimates of biases, estimates of positions and proprioceptive measurements) reduce significantly systematic errors. However, cooperative localization based on sharing estimates is subject to data incest problems (i.e, reuse of identical information in the fusion process) that often lead to over-convergence problems. When position information is used in a safety-critical context (e.g. close navigation of autonomous robots), one should check the consistency of the localization estimates. In this context, we aim at characterizing reliable confidence domains that contain robots positions with high reliability. Hence, set-membership methods are considered as efficient solutions. This kind of approach enables merging adequately the information even when it is reused several time. It also provides reliable domains. Moreover, the use of non-linear models does not require any linearization. The modeling of a cooperative system of nr robots with biased beacons measurements is firstly presented. Then, we perform an observability study. Two cases regarding the localization technology are considered. Observability conditions are identified and demonstrated. We then propose a set-membership method for cooperativelocalization. Cooperation is performed by sharing estimated positions, estimated biases and proprioceptive measurements. Sharing biases estimates allows to reduce the estimation error and the uncertainty of the robots positions. The algorithm feasibility is validated through simulation when the observations are beacons distance measurements with several robots. The cooperation provides better performance compared to a non-cooperative method. Afterwards, the cooperative algorithm based on set-membership method is tested using real data with two experimental vehicles. Finally, we compare the interval method performance with a sequential Bayesian approach based on covariance intersection. Experimental results indicate that the interval approach provides more accurate positions of the vehicles with smaller confidence domains that remain reliable. Indeed, the comparison is performed in terms of accuracy and uncertainty.

L’objectif de cette thèse est de développer des méthodes permettant la cartographie d’un environnement tridimensionnel de grande dimension en combinant radar panoramique MMW et caméras optiques. Contrairement aux méthodes existantes de fusion de données multi-capteurs, telles que le SLAM, nous souhaitons réaliser un capteur de type RGB-D fournissant directement des mesures de profondeur enrichies par l’apparence (couleur, texture...). Après avoir modélisé géométriquement le système radar/caméra, nous proposons une méthode de calibrage originale utilisant des correspondances de points. Pour obtenir ces correspondances, des cibles permettant une mesure ponctuelle aussi bien par le radar que la caméra ont été conçues. L’approche proposée a été élaborée pour pouvoir être mise en oeuvre dans un environnement libre et par un opérateur non expert. Deuxièmement, une méthode de reconstruction de points tridimensionnels sur la base de correspondances de points radar et image a été développée. Nous montrons par une analyse théorique des incertitudes combinées des deux capteurs et par des résultats expérimentaux, que la méthode proposée est plus précise que la triangulation stéréoscopique classique pour des points éloignés comme on en trouve dans le cas de cartographie d’environnements extérieurs. Enfin, nous proposons une stratégie efficace de mise en correspondance automatique des données caméra et radar. Cette stratégie utilise deux caméras calibrées. Prenant en compte l’hétérogénéité des données radar et caméras, l’algorithme développé commence par segmenter les données radar en régions polygonales. Grâce au calibrage, l’enveloppe de chaque région est projetée dans deux images afin de définir des régions d’intérêt plus restreintes. Ces régions sont alors segmentées à leur tour en régions polygonales générant ainsi une liste restreinte d’appariement candidats. Un critère basé sur l’inter corrélation et la contrainte épipolaire est appliqué pour valider ou rejeter des paires de régions. Tant que ce critère n’est pas vérifié, les régions sont, elles même, subdivisées par segmentation. Ce processus, favorise l’appariement de régions de grande dimension en premier. L’objectif de cette approche est d’obtenir une cartographie sous forme de patchs localement denses. Les méthodes proposées, ont été testées aussi bien sur des données de synthèse que sur des données expérimentales réelles. Les résultats sont encourageants et montrent, à notre sens, la faisabilité de l’utilisation de ces deux capteurs pour la cartographie d’environnements extérieurs de grande échelle. / The main goal of this PhD work is to develop 3D mapping methods of large scale environment by combining panoramic radar and cameras. Unlike existing sensor fusion methods, such as SLAM (simultaneous localization and mapping), we want to build a RGB-D sensor which directly provides depth measurement enhanced with texture and color information. After modeling the geometry of the radar/camera system, we propose a novel calibration method using points correspondences. To obtain these points correspondences, we designed special targets allowing accurate point detection by both the radar and the camera. The proposed approach has been developed to be implemented by non-expert operators and in unconstrained environment. Secondly, a 3D reconstruction method is elaborated based on radar data and image point correspondences. A theoretical analysis is done to study the influence of the uncertainty zone of each sensor on the reconstruction method. This theoretical study, together with the experimental results, show that the proposed method outperforms the conventional stereoscopic triangulation for large scale outdoor scenes. Finally, we propose an efficient strategy for automatic data matching. This strategy uses two calibrated cameras. Taking into account the heterogeneity of cameras and radar data, the developed algorithm starts by segmenting the radar data into polygonal regions. The calibration process allows the restriction of the search by defining a region of interest in the pair of images. A similarity criterion based on both cross correlation and epipolar constraint is applied in order to validate or reject region pairs. While the similarity test is not met, the image regions are re-segmented iteratively into polygonal regions, generating thereby a shortlist of candidate matches. This process promotes the matching of large regions first which allows obtaining maps with locally dense patches. The proposed methods were tested on both synthetic and real experimental data. The results are encouraging and prove the feasibility of radar and vision sensor fusion for the 3D mapping of large scale urban environment.

L'objectif de la thèse est de développer des algorithmes de fusion de données recueillies à l’aide d'un réseau de capteurs sans fil afin de localiser plusieurs sources émettant un agent chimique ou biologique dans l'air. Ces capteurs détectent la concentration de la substance émise, transportée par advection et diffusion, au niveau de leurs positions et de communiquer cette information à un centre de traitement. L’information recueillie de façon collaborative est d'abord utilisée pour localiser les capteurs déployés au hasard et ensuite pour localiser les sources. Les applications comprennent, entre autres, la surveillance environnementale et la surveillance de sites sensibles ainsi que des applications de sécurité dans le cas d'une libération accidentelle ou intentionnelle d'un agent toxique. Toutefois, l'application considérée dans la thèse est celle de la détection et la localisation de mines terrestres. Dans cette approche, les mines sont considérées comme des sources émettrices de produits chimiques explosifs.La thèse comprend une contribution théorique où nous étendons l'algorithme de propagation de la croyance, un algorithme de fusion de données bien connu et largement utilisé pour l'estimation collaborative d'état dans les réseaux de capteurs, au cadre des méthodes à erreurs bornées. Le nouvel algorithme est testé sur le problème de l'auto-localisation dans les réseaux de capteurs statiques ainsi que l'application de suivi d'un objet mobile en utilisant un réseau de capteurs de distance. Autres contributions comprennent l'utilisation d'une approche probabiliste bayésienne avec des techniques d'analyse de données pour localiser un nombre inconnu de sources émettrices de vapeur. / The aim of the thesis is to develop fusion algorithms for data collected from a wireless sensor network in order to locate multiple sources emitting some chemical or biological agent in the air. These sensors detect the concentration of the emitted substance, transported by advection and diffusion, at their positions and communicate this information to a treatment center. The information collected in a collaborative manner is used first to locate the randomly deployed sensors and second to locate the sources. Applications include, amongst others, environmental monitoring and surveillance of sensitive sites as well as security applications in the case of an accidental or intentional release of a toxic agent. However, the application we consider in the thesis is that of landmine detection and localization. In this approach, the land mines are considered as sources emitting explosive chemicals. The thesis includes a theoretical contribution where we extend the Belief Propagation algorithm, a well-known data fusion algorithm that is widely used for collaborative state estimation in sensor networks, to the bounded error framework. The novel algorithm is tested on the self-localization problem in static sensor networks as well as the application of tracking a mobile object using a network of range sensors. Other contributions include the use of a Bayesian probabilistic approach along with data analysis techniques to locate an unknown number of vapor emitting sources.

En France, les chutes constituent la première cause de mortalité chez les plus de 75 ans, et la seconde chez les plus de 65 ans. On estime qu'elle engendre un coût de 1 à 2 milliards d'euros par an pour la société. L'enjeu humain et socio-économique est colossal, sachant que le risque de chute est multiplié par 20 après une première chute, que le risque de décès est multiplié par 4 dans l'année qui suit une chute, que les chutes concernent 30% des personnes de plus de 65 ans et 50% des personnes de plus de 85 ans, et que l'on estime que d'ici 2050, plus de 30% de la population sera âgée de plus de 65 ans. Cette thèse propose un dispositif de détection de présence au sol se basant sur l'analyse de cartes de profondeurs acquises en temps réel, ainsi qu'une amélioration du dispositif proposé utilisant également un capteur thermique. Les cartes de profondeurs et les images thermiques nous permettent de nous affranchir des conditions d'illumination de la scène observée, et garantissent l'anonymat des personnes qui évoluent dans le champ de vision du dispositif. Cette thèse propose également différentes méthodes de détection du plan du sol dans une carte de profondeurs, le plan du sol constituant une référence géométrique nécessaire au dispositif proposé. Une enquête psychosociale a été réalisée, qui nous a permis d'évaluer l'acceptabilité a priori dudit dispositif. Cette enquête a démontré sa bonne acceptabilité, et a fourni des préconisations quant aux points d'amélioration et aux écueils à éviter. Enfin, une méthode de suivi d'objets dans une carte de profondeurs est proposée, un objectif à plus long terme consistant à mesurer l'activité des individus observés. / In France, fall is the first death cause for people aged 75 and more, and the second death cause for people aged 65 and more. It is considered that falls generate about 1 to 2 billion euros health costs per year. The human and social-economical issue is crucial, knowing that for the mentioned populations, fall risk is multiplied by 20 after a first fall; that the death risk is multiplied by 4 in the year following a fall; that per year, 30% of the people aged 65 and more and 50% of the people aged 85 and more are subject to falls; and that it is estimated that more than 30% of the French population whill be older than 65 years old by 2050. This thesis proposes a ground lying event detection device which bases on the real time analysis of depth maps, and also proposes an improvement of the device, which uses an additional thermal sensor. Depth maps and thermal images ensure the device is independent from textures and lighting conditions of the observed scenes, and guarantee that the device respects the privacy of those who pass into its field of view, since nobody can be recognized in such images. This thesis also proposes several methods to detect the ground plane in a depth map, the ground plane being a geometrical reference for the device. A psycho-social inquiry was conducted, and enabled the evaluation of the a priori acceptability of the proposed device. This inquiry demonstrated the good acceptability of the proposed device, and resulted in recommendations on points to be improved and on pitfalls to avoid. Last, a method to separate and track objects detected in a depth map is proposed, the measurement of the activity of observed individuals being a long term objective for the device.