Design and implementation of a relative localization system for ground and aerial robotic teams

The main focus of this thesis is to address the relative localization problem of a heterogenous team which comprises of both ground and micro aerial vehicle robots. This team configuration allows to combine the advantages of increased accessibility and better perspective provided by aerial robots with the higher computational and sensory resources provided by the ground agents, to realize a cooperative multi robotic system suitable for hostile autonomous missions. However, in such a scenario, the strict constraints in flight time, sensor pay load, and computational capability of micro aerial vehicles limits the practical applicability of popular map-based localization schemes for GPS denied navigation. Therefore, the resource limited aerial platforms of this team demand simpler localization means for autonomous navigation. Relative localization is the process of estimating the formation of a robot team using the acquired inter-robot relative measurements. This allows the team members to know their relative formation even without a global localization reference, such as GPS or a map. Thus a typical robot team would benefit from a relative localization service since it would allow the team to implement formation control, collision avoidance, and supervisory control tasks, independent of a global localization service. More importantly, a heterogenous team such as ground robots and computationally constrained aerial vehicles would benefit from a relative localization service since it provides the crucial localization information required for autonomous operation of the weaker agents. This enables less capable robots to assume supportive roles and contribute to the more powerful robots executing the mission. Hence this study proposes a relative localization-based approach for ground and micro aerial vehicle cooperation, and develops inter-robot measurement, filtering, and distributed computing modules, necessary to realize the system. The research study results in three significant contributions. First, the work designs and validates a novel inter-robot relative measurement hardware solution which has accuracy, range, and scalability characteristics, necessary for relative localization. Second, the research work performs an analysis and design of a novel nonlinear filtering method, which allows the implementation of relative localization modules and attitude reference filters on low cost devices with optimal tuning parameters. Third, this work designs and validates a novel distributed relative localization approach, which harnesses the distributed computing capability of the team to minimize communication requirements, achieve consistent estimation, and enable efficient data correspondence within the network. The work validates the complete relative localization-based system through multiple indoor experiments and numerical simulations. The relative localization based navigation concept with its sensing, filtering, and distributed computing methods introduced in this thesis complements system limitations of a ground and micro aerial vehicle team, and also targets hostile environmental conditions. Thus the work constitutes an essential step towards realizing autonomous navigation of heterogenous teams in real world applications.

Sensor Planning for Bayesian Nonparametric Target Modeling

Bayesian nonparametric models, such as the Gaussian process and the Dirichlet process, have been extensively applied for target kinematics modeling in various applications including environmental monitoring, traffic planning, endangered species tracking, dynamic scene analysis, autonomous robot navigation, and human motion modeling. As shown by these successful applications, Bayesian nonparametric models are able to adjust their complexities adaptively from data as necessary, and are resistant to overfitting or underfitting. However, most existing works assume that the sensor measurements used to learn the Bayesian nonparametric target kinematics models are obtained a priori or that the target kinematics can be measured by the sensor at any given time throughout the task. Little work has been done for controlling the sensor with bounded field of view to obtain measurements of mobile targets that are most informative for reducing the uncertainty of the Bayesian nonparametric models. To present the systematic sensor planning approach to leaning Bayesian nonparametric models, the Gaussian process target kinematics model is introduced at first, which is capable of describing time-invariant spatial phenomena, such as ocean currents, temperature distributions and wind velocity fields. The Dirichlet process-Gaussian process target kinematics model is subsequently discussed for modeling mixture of mobile targets, such as pedestrian motion patterns.

Novel information theoretic functions are developed for these introduced Bayesian nonparametric target kinematics models to represent the expected utility of measurements as a function of sensor control inputs and random environmental variables. A Gaussian process expected Kullback Leibler divergence is developed as the expectation of the KL divergence between the current (prior) and posterior Gaussian process target kinematics models with respect to the future measurements. Then, this approach is extended to develop a new information value function that can be used to estimate target kinematics described by a Dirichlet process-Gaussian process mixture model. A theorem is proposed that shows the novel information theoretic functions are bounded. Based on this theorem, efficient estimators of the new information theoretic functions are designed, which are proved to be unbiased with the variance of the resultant approximation error decreasing linearly as the number of samples increases. Computational complexities for optimizing the novel information theoretic functions under sensor dynamics constraints are studied, and are proved to be NP-hard. A cumulative lower bound is then proposed to reduce the computational complexity to polynomial time.

Three sensor planning algorithms are developed according to the assumptions on the target kinematics and the sensor dynamics. For problems where the control space of the sensor is discrete, a greedy algorithm is proposed. The efficiency of the greedy algorithm is demonstrated by a numerical experiment with data of ocean currents obtained by moored buoys. A sweep line algorithm is developed for applications where the sensor control space is continuous and unconstrained. Synthetic simulations as well as physical experiments with ground robots and a surveillance camera are conducted to evaluate the performance of the sweep line algorithm. Moreover, a lexicographic algorithm is designed based on the cumulative lower bound of the novel information theoretic functions, for the scenario where the sensor dynamics are constrained. Numerical experiments with real data collected from indoor pedestrians by a commercial pan-tilt camera are performed to examine the lexicographic algorithm. Results from both the numerical simulations and the physical experiments show that the three sensor planning algorithms proposed in this dissertation based on the novel information theoretic functions are superior at learning the target kinematics with

little or no prior knowledge

Mechanochemistry for Active Materials and Devices

The coupling of mechanical stress fields in polymers to covalent chemistry (polymer mechanochemistry) has provided access to previously unattainable chemical reactions and polymer transformations. In the bulk, mechanochemical activation has been used as the basis for new classes of stress-responsive polymers that demonstrate stress/strain sensing, shear-induced intermolecular reactivity for molecular level remodeling and self-strengthening, and the release of acids and other small molecules that are potentially capable of triggering further chemical response. The potential utility of polymer mechanochemistry in functional materials is limited, however, by the fact that to date, all reported covalent activation in the bulk occurs in concert with plastic yield and deformation, so that the structure of the activated object is vastly different from its nascent form. Mechanochemically activated materials have thus been limited to “single use” demonstrations, rather than as multi-functional materials for structural and/or device applications. Here, we report that filled polydimethylsiloxane (PDMS) elastomers provide a robust elastic substrate into which mechanophores can be embedded and activated under conditions from which the sample regains its original shape and properties. Fabrication is straightforward and easily accessible, providing access for the first time to objects and devices that either release or reversibly activate chemical functionality over hundreds of loading cycles.

While the mechanically accelerated ring-opening reaction of spiropyran to merocyanine and associated color change provides a useful method by which to image the molecular scale stress/strain distribution within a polymer, the magnitude of the forces necessary for activation had yet to be quantified. Here, we report single molecule force spectroscopy studies of two spiropyran isomers. Ring opening on the timescale of tens of milliseconds is found to require forces of ~240 pN, well below that of previously characterized covalent mechanophores. The lower threshold force is a combination of a low force-free activation energy and the fact that the change in rate with force (activation length) of each isomer is greater than that inferred in other systems. Importantly, quantifying the magnitude of forces required to activate individual spiropyran-based force-probes enables the probe behave as a “scout” of molecular forces in materials; the observed behavior of which can be extrapolated to predict the reactivity of potential mechanophores within a given material and deformation.

We subsequently translated the design platform to existing dynamic soft technologies to fabricate the first mechanochemically responsive devices; first, by remotely inducing dielectric patterning of an elastic substrate to produce assorted fluorescent patterns in concert with topological changes; and second, by adopting a soft robotic platform to produce a color change from the strains inherent to pneumatically actuated robotic motion. Shown herein, covalent polymer mechanochemistry provides a viable mechanism to convert the same mechanical potential energy used for actuation into value-added, constructive covalent chemical responses. The color change associated with actuation suggests opportunities for not only new color changing or camouflaging strategies, but also the possibility for simultaneous activation of latent chemistry (e.g., release of small molecules, change in mechanical properties, activation of catalysts, etc.) in soft robots. In addition, mechanochromic stress mapping in a functional actuating device might provide a useful design and optimization tool, revealing spatial and temporal force evolution within the actuator in a way that might also be coupled to feedback loops that allow autonomous, self-regulation of activity.

In the future, both the specific material and the general approach should be useful in enriching the responsive functionality of soft elastomeric materials and devices. We anticipate the development of new mechanophores that, like the materials, are reversibly and repeatedly activated, expanding the capabilities of soft, active devices and further permitting dynamic control over chemical reactivity that is otherwise inaccessible, each in response to a single remote signal.

Intraspecific Behavioural Diversity in a Freshwater Zooplankton: A study of the fitness consequences of vertical migration behaviour for distinct morphs of Daphnia pulicaria

The literature on niche separation and coexistence between species is large, but there is widespread variation in behavioural strategy between individuals of the same species that has received much less attention. Understanding what maintains this diversity is important because intraspecific behavioural diversity can affect population dynamics and community interactions. Multiple behavioural strategies can arise either as phenotype-dependent ‘conditional strategies’, where phenotypic variation causes individuals to adopt different strategies for optimizing fitness, or as internally-independent ‘alternative strategies’, where multiple fitness peaks exist for individuals and strategic ‘choice’ remains plastic. Though intraspecific variation in stable phenotypes is known to maintain intraspecific behavioural diversity through conditional strategies, when internal conditions are highly plastic or reversible, it is not clear whether individual behaviours are maintained as conditional strategies, or as alternative strategies of equal fitness. In this study, I combine an observational and experimental approach to identify the likely mechanisms maintaining behavioural diversity between hemoglobin-rich and hemoglobin-poor morphs in a natural population of Daphnia pulicaria. In Round Lake, individuals with low hemoglobin migrate daily from the hypolimnion to the epilimnion, whereas individuals with high hemoglobin remain in the hypolimnion. Using high-resolution depth and time sampling, I discovered behavioural diversity both within and among hemoglobin phenotypes. I tested the role of hemoglobin phenotype in maintaining behavioural diversity using automated migration robots that move individuals across the natural environmental gradients in the lake. By measuring the fitness of each morph undergoing either a natural migration behaviour, or the migration of the opposite morph, I found that the fitness of hemoglobin rich and poor morphs in their natural behaviour does not differ, but that Hb-rich individuals can obtain equal fitness from either behaviour, while Hb-poor morphs suffer substantial drops in survivorship in the alternate migration behaviour. Thus, migration behaviour in this system exists as a conditional strategy for some individuals, and as alternative strategies of equal fitness for others. The results of this study suggest that individual limits in the expression of highly flexible internal conditions can reinforce intraspecific behavioural diversity. Few studies have measured the fitness consequences of switching migration strategies and this study provides a rare example in the field.

News agencies, their constant reinvention as a strategy to face competition

News agencies compete for a foothold as providers of information and mass media. Covered by a technological class infrastructure, Associated Press, Reuters, Agence FrancePresse (AFP) and EFE are leaders of the global media system because they introduce revolutionary changes in their production routines, professional culture, journalistic genres and styles; also for its innovative product offerings and services. This article also focuses on the strategies of the agencies to get closer to their audiences, from the agreements established and the treatment of very specific themes. Some solutions that contribute to the future survival of these entities are also proposed.

Multimodal attention system for an interactive robot

[EN]Social robots are receiving much interest in the robotics community. The most important goal for such robots lies in their interaction capabilities. An attention system is crucial, both as a filter to center the robot’s perceptual resources and as a mean of letting the observer know that the robot has intentionality. In this paper a simple but flexible and functional attentional model is described. The model, which has been implemented in an interactive robot currently under development, fuses both visual and auditive information extracted from the robot’s environment, and can incorporate knowledge-based influences on attention.

Eldi: an agent based museum robot

[EN]In this paper we will present Eldi, a mobile robot that has been in daily operation at the Elder Museum of Science and Technology at Las Palmas de Gran Canaria since December 1999. This is an ongoing project that was organized in three di erent stages, describing here the one that has been accomplished. The initial phase, termed \The Player", the second stage, actually under development, has been called "The Cicerone" and in a nal phase, termed \The Vagabond", Eldi will be allowed to move erratically across the Museum. This paper will focus on the accomplished rst stage to succinctly describe the physical robot and the environment and demos developed. Finally we will summarize some important lessons learnt.

Experiences with a museum robot

In this paper we will present Eldi, a mobile robot that has been in daily operation at the Elder Museum of S ien e and Te hnology at Las Palmas de Gran Canaria sin e last De ember. This is an ongoing pro je t that was organized in three di erent stages of whi h only the rst one has been a omplished. The initial phase, termed \The Player", the se ond stage, a tually under development, has been alled "The Ci erone" and in a nal phase, termed \The Vagabond", Eldi will be allowed to move errati ally a ross the Museum...

An active vision system integrating fast and slow processes

[EN]This paper describes an Active Vision System whose design assumes a distinction between fast or reactive and slow or background processes. Fast processes need to operate in cycles with critical timeouts that may affect system stability. While slow processes, though necessary, do not compromise system stability if its execution is delayed. Based on this simple taxonomy, a control architecture has been proposed and a prototype implemented that is able to track people in real-time with a robotic head while trying to identify the target. In this system, the tracking module is considered as the reactive part of the system while person identification is considered a background task.

Localization of a mobile robot based on odometry and natural landmarks using extended kalman filter

SANTANA, André M.; SOUZA, Anderson A. S.; BRITTO, Ricardo S.; ALSINA, Pablo J.; MEDEIROS, Adelardo A. D. Localization of a mobile robot based on odometry and natural landmarks using extended Kalman Filter. In: INTERNATIONAL CONFERENCE ON INFORMATICS IN CONTROL, AUTOMATION AND ROBOTICS, 5., 2008, Funchal, Portugal. Proceedings... Funchal, Portugal: ICINCO, 2008.

II Taller de Innovación Educativa de la Escuela de Ingenierías Industriales y Civiles

Videos de las presentaciones efectuadas en el II Taller de Innovación Educativa de la EIIC. Donde se pretende conocer las experiencias docentes innovadoras llevadas a cabo por profesores tanto de la Escuela como de la Universidad. Ponencias: Experiencias en el desarrollo de materiales didáctivos... / P.M. Hernández Castellano; Aprendizaje por proyectos utilizando tecnologías y robots.../ Daura Vega Moreno; El repositorio institucional Acceda... / Avelina Fernández; Experiencias en el uso de OpenULPGC en la formación on-line/ Luis Álvarez; UBILINGUA para un aprendizaje IFE de la ULPGC/ Soraya García Sänchez y Carmen Luján; Análisis de la opinión de los estudiantes sobre la aplicación de trabajo colaborativo.../ J.M. Herrera Melián.

Diseño, construcción y control de un robot balancín.

Se diseñó, construyó, simuló e implementó un dispositivo robot balancín para la aplicación y estudio de técnicas avanzadas de control. Para esto se realizó el diseño mecánico del dispositivo, de acuerdo a una elección entre dos modelos distintos y cuatro tipos diferentes de transmisión. Luego se instrumentó el dispositivo con encoders de posición , acelerómetro y giróscopo para obtener el estado del dispositivo y controlarlo. Se realizó una placa electrónica para la lectura y procesamiento de señales de sensores con un micro controlador, un regulador de tensión, y un driver para los motores, capaz de obtener las señales de los encoders y el módulo acelerómetro-giróscopo y enviarlas por comunicación hacia una mini-computadora, la cual ejecuta el control, y se comunica nuevamente a la placa diseñada para comandar los motores. Se desarrolló un modelo teórico simplificado en dos dimensiones para facilitar la posterior identificación de planta. Se realizaron experimentos para lograr una identificación de planta. A partir de lo obtenido, se diseñó y simuló el control necesario para mantener la estabilidad. Se implementó posteriormente el control diseñado. Se reajustaron los parámetros correspondientes de acuerdo a la práctica experimental para mejorar la respuesta dinámica del sistema.

Développement d'algorithmes d'estimation de la pose d'objets saisis par un préhenseur robotique

Les préhenseurs robotiques sont largement utilisés en industrie et leur déploiement pourrait être encore plus important si ces derniers étaient plus intelligents. En leur conférant des capacités tactiles et une intelligence leur permettant d’estimer la pose d’un objet saisi, une plus vaste gamme de tâches pourraient être accomplies par les robots. Ce mémoire présente le développement d’algorithmes d’estimation de la pose d’objets saisis par un préhenseur robotique. Des algorithmes ont été développés pour trois systèmes robotisés différents, mais pour les mêmes considérations. Effectivement, pour les trois systèmes la pose est estimée uniquement à partir d’une saisie d’objet, de données tactiles et de la configuration du préhenseur. Pour chaque système, la performance atteignable pour le système minimaliste étudié est évaluée. Dans ce mémoire, les concepts généraux sur l’estimation de la pose sont d’abord exposés. Ensuite, un préhenseur plan à deux doigts comprenant deux phalanges chacun est modélisé dans un environnement de simulation et un algorithme permettant d’estimer la pose d’un objet saisi par le préhenseur est décrit. Cet algorithme est basé sur les arbres d’interprétation et l’algorithme de RANSAC. Par la suite, un système expérimental plan comprenant une phalange supplémentaire par doigt est modélisé et étudié pour le développement d’un algorithme approprié d’estimation de la pose. Les principes de ce dernier sont similaires au premier algorithme, mais les capteurs compris dans le système sont moins précis et des adaptations et améliorations ont dû être appliquées. Entre autres, les mesures des capteurs ont été mieux exploitées. Finalement, un système expérimental spatial composé de trois doigts comprenant trois phalanges chacun est étudié. Suite à la modélisation, l’algorithme développé pour ce système complexe est présenté. Des hypothèses partiellement aléatoires sont générées, complétées, puis évaluées. L’étape d’évaluation fait notamment appel à l’algorithme de Levenberg-Marquardt.

Développement d’un système d’actionnement utilisant la combustion d’une source d’énergie chimique pour la robotique mobile

Les systèmes d’actionnement couramment utilisés sur les systèmes de robotiques mobiles tels que les exosquelettes ou les robots marcheurs sont majoritairement électriques. Les batteries couplées à des moteurs électriques souffrent toutefois d’une faible densité de stockage énergétique et une faible densité de puissance, ce qui limite l’autonomie de ces dispositifs pour une masse de système donnée. Une étude comparative des systèmes d’actionnement potentiels a permis de déterminer que l’utilisation d’une source d’énergie chimique permettait d’obtenir une densité de stockage énergétique supérieure aux batteries. De plus, il a été déterminé que l’utilisation de la combustion directement dans un actionneur pneumatique souple permettrait d’obtenir une densité de puissance beaucoup plus élevée que celle des moteurs électriques. La conception, la fabrication et la caractérisation de plusieurs types d’actionneurs pneumatiques pressurisés directement par la combustion d’une source d’énergie chimique ont permis d’évaluer la faisabilité de l’approche, dans un contexte de robotique mobile, plus précisément pour des tâches de locomotion. Les paramètres permettant d’obtenir une efficacité énergétique élevée ont été étudiés. Il a été démontré que le ratio de compression et le ratio d’expansion doivent être optimisés. De plus, comme les pertes thermiques sont le mécanisme de perte dominant, le ratio d’équivalence devrait être réduit au minimum. Parmi les carburants usuels, l’hydrogène permet d’atteindre les valeurs de ratio d’équivalence les plus faibles, ce qui en fait un choix de carburant idéal. Les résultats expérimentaux ont été utilisés pour corréler un modèle analytique d’un actionneur pneumatique à combustion. Ce modèle analytique est par la suite utilisé pour vérifier la faisabilité théorique de l’utilisation de l’approche d’actionnement pour fournir la puissance à un dispositif d’assistance à la locomotion.