<p dir="ltr">State estimation is a type of an <i>inverse problem</i> in which some amount of observed data needs to be processed using computer algorithms (which are designed using analytical techniques) to infer or reconstruct the underlying model that produced the data. Due to the ubiquity of data and interconnected control systems in the present day, many engineering domains have become replete with inverse problems that can be formulated as state estimation problems. The interconnectedness of these control systems imparts the associated state estimation problems with distinctive structural properties that must be taken into consideration. For instance, the observed data could be high-dimensional and have a dependency structure that is best described by a graph. Furthermore, the control systems of today interface with each other and with the internet, bringing in new possibilities for large-scale collaborative sensor fusion, while also (potentially) introducing new sources of disturbances, faults, and cyberattacks. </p><p dir="ltr">The main thesis of this document is to investigate the unique challenges related to the issues of robustness, resilience (to faults and cyberattacks), and scalability of state estimation algorithms. These correspond to research questions such as, <i>"Does the state estimation algorithm retain its performance when the measurements are perturbed by unknown disturbances or adversarial inputs?"</i> and <i>"Does the algorithm have any bottlenecks that restrict the size/dimension of the problems that it could be applied to?".</i> Most of these research questions are motivated by a singular domain of application: autonomous navigation of unmanned aerial vehicles (UAVs). Nevertheless, the mathematical methods and research philosophy employed herein are quite general, making the results of this document applicable to a variety of engineering tasks, including anomaly detection in time-series data, autonomous remote sensing, traffic monitoring, coordinated motion of dynamical systems, and fault-diagnosis of wireless sensor networks (WSNs), among others.</p>

Society today brings a high pace development and demand of Artificial intelligence systems as well as robotics. To further expand and to take one step closer to have Unmanned Aerial Vehicles (UAVs) working in the cities, the European Union Aviation Safety Agency launched a project that introduces U-space airspace, an airspace where UAVs, for instance, are allowed to operate for commercial services.The problems defined for U-space airspace resemble problems defined in the area of multi-agent path finding, such as scaling and traffic etc., resulting an interest to research whether MAPF-solutions can be applied to U-space scenarios. The following thesis extends the state-of-the-art MAPF-algorithm Continuous-time Conflict based search (CCBS) to handle simplified U-space scenarios, as well as extend other A*-based algorithms, such as a version of the Receding Horizon Lattice-based Motion Planning named Extended Multi-agent A* algorithm with Wait-Time (EMAWT) and an extended A* named Extended Multi-agent A* algorithm (EMA) to handle them. Comparisons of the three algorithms resulted in the EMAWT being the most reliable and stable solution throughout all tests, whilst for fewer agents, the CCBS being the clear best solution.

No description available.

Multi-agent systems are employed for a group of agents to achieve coordinated tasks, in which distributed sensing, computing, communication and control are usually integrated with shared resources. Efficient usage of these resources is therefore an important issue. In addition, in applications such as robotics, a group of agents may encounter the request of a sequence of tasks and deadline constraint on the completion of each task is a common requirement. Thus, the integration of multi-agent task scheduling and control synthesis is of great practical interest. In this thesis, we study control of multi-agent systems under a networked control system framework. The first purpose is to design resource-efficient communication and control strategies to solve consensus problem for multi-agent systems.The second purpose is to jointly schedule task sequence and design controllers for multiagent systems that are subject to a sequence of deadline-constrained tasks. In the first part, a distributed asynchronous event-triggered communication and control strategy is proposed to tackle multi-agent consensus. It is shown that the proposed event-triggered communication and control strategy fulfils the reduction of both the rates of sensor-controller communication and controller-actuator communication as well as excluding Zeno behavior. To further relax the requirement of continuous sensing and computing, a periodic event-triggered communication and control strategy is proposed in the second part. In addition, an observer-based encoder-decoder with finite-level quantizeris designed to deal with the constraint of limited data rate. An explicit formula for the maximum allowable sampling period is derived first. Then, it is proven that exponential consensus can be achieved in the presence of data rate constraint. Finally, in the third part, the problem of deadline-constrained multi-agent task scheduling and control synthesis is addressed. A dynamic scheduling strategy is proposed and a distributed hybrid control law is designed for each agent that guarantees the completion and deadline satisfaction of each task. The effectiveness of the theoretical results in the thesis is verified by several simulation examples. / <p>QC 20180918</p>

Formation control is a classical problem and has been a prime topic of interest among the scientific community in the past few years. Although a vast amount of literature exists in this field, there are still many open questions that require an in-depth understanding and a new perspective. This thesis contributes towards exploring the wide dimensions of formation control and implementing a formation control scheme for a group of multi-agent systems. These systems are autonomous in nature and are represented by double integrated dynamics. It is assumed that the agents are connected in an undirected graph and use a leader-follower architecture to reach formation when the leading agent is given a velocity that is piecewise constant. A MATLAB code is written for the implementation of formation and the consensus-based control laws are verified. Understanding the effects on formation due to a fixed formation geometry is also observed and reported. Also, a link that describes the functional similarity between desired formation geometry and the Laplacian matrix has been observed. The use of Laplacian matrix in stability analysis of the formation is of special interest.

La pérdida de privacidad se está convirtiendo en uno de los mayores problemas en el mundo de la informática. De hecho, la mayoría de los usuarios de Internet (que hoy en día alcanzan la cantidad de 2 billones de usuarios en todo el mundo) están preocupados por su privacidad. Estas preocupaciones también se trasladan a las nuevas ramas de la informática que están emergiendo en los ultimos años. En concreto, en esta tesis nos centramos en la privacidad en los Sistemas Multiagente. En estos sistemas, varios agentes (que pueden ser inteligentes y/o autónomos) interactúan para resolver problemas. Estos agentes suelen encapsular información personal de los usuarios a los que representan (nombres, preferencias, tarjetas de crédito, roles, etc.). Además, estos agentes suelen intercambiar dicha información cuando interactúan entre ellos. Todo esto puede resultar en pérdida de privacidad para los usuarios, y por tanto, provocar que los usuarios se muestren adversos a utilizar estas tecnologías. En esta tesis nos centramos en evitar la colección y el procesado de información personal en Sistemas Multiagente. Para evitar la colección de información, proponemos un modelo para que un agente sea capaz de decidir qué atributos (de la información personal que tiene sobre el usuario al que representa) revelar a otros agentes. Además, proporcionamos una infraestructura de agentes segura, para que una vez que un agente decide revelar un atributo a otro, sólo este último sea capaz de tener acceso a ese atributo, evitando que terceras partes puedan acceder a dicho atributo. Para evitar el procesado de información personal proponemos un modelo de gestión de las identidades de los agentes. Este modelo permite a los agentes la utilización de diferentes identidades para reducir el riesgo del procesado de información. Además, también describimos en esta tesis la implementación de dicho modelo en una plataforma de agentes. / Such Aparicio, JM. (2011). ENHANCING PRIVACY IN MULTI-AGENT SYSTEMS [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/13023

Actualmente, los sistemas informáticos complejos se describen en términos de entidades que actúan como proveedores y consumidores. Estas entidades ofrecen su funcionalidad a través de servicios e interactúan entre ellas para ofrecer o pedir estos servicios. La integración de Sistemas Multi-Agente Abiertos y de Sistemas Orientados a Servicios es adecuada para implementar este tipo de sistemas. En los SMA abiertos, los agentes entran y salen del sistema, interactúan con los demás de una manera flexible, y se consideran como entidades reactivas y proactivas, capaces de razonar acerca de lo que sucede en su entorno y llevar a cabo acciones locales sobre la base de sus observaciones para alcanzar sus metas. El área de la computación orientada a servicios proporciona los bloques de construcción básicos para aplicaciones empresariales complejas que son los servicios. Los servicios son independientes de la plataforma y pueden ser descubiertos y compuestos de manera dinámica. Estas características hacen que los servicios sean adecuados para hacer frente a la elevada tasa de cambios en las demandas de las empresas. Sin embargo, la complejidad de los sistemas informáticos, los cambios en las condiciones del entorno y el conocimiento parcial de los agentes sobre el sistema requieren que los agentes cuenten con mecanismos que les faciliten tareas como el descubrimiento de servicios, la auto-organización de sus relaciones estructurales conforme se producen cambios en la demanda de servicios, y la promoción y mantenimiento de un comportamiento cooperativo entre los agentes para garantizar el buen desarrollo de la actividad de descubrimiento de servicios en el sistema. La principal aportación de esta tesis doctoral es la propuesta de un marco para Sistemas Multi-Agente Abiertos Orientados a Servicios. Este marco integra agentes que se encuentran en una red sin ningún tipo de estructura predefinida, y agentes que además de estar en esa red forman parte de grupos dinámicos más comp / Del Val Noguera, E. (2013). Semantic Service management for service-oriented MAS [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/27556

Organizations have become the backbone of the society. Humans live around all kinds of organizations, such as neighborhood communities, businesses, schools, unions, political, sports, and religious organizations, etc. These organizations have a set of members, each playing a specific role, which determines their duties and functionalities within the organization. One of these functionalities is to offer a range of services to members of the organization and external people. These members must follow a set of norms to ensure the proper functioning of the organization and should pursue the global goals of the organization. A feature that is repeated in organizations is that they are not static but dynamic, resulting in changes in both its structure and the way in which they behave. In an organization, any of its elements is prone to change due to situations that occur in the organization itself or its environment. Researchers in the field of social sciences and organizations have studied such situations, the reasons why they appear and solutions and actions to be taken to ensure that this situation does not damage the organization or to take advantage of the situation. These situations are known as ‘Forces that drive organizational change’. Human organizations are the main source of inspiration for the Multi-Agent Systems (MAS) based on organizations. These systems are computational abstractions that are populated by agents instead of people, but take into account organizational elements such as roles, services, goals, norms, etc. However, the proposals that have been presented up to now to define this type of MAS are focused mostly on static systems, without changes in its structure. Moreover, in the few proposals that take into account organizational changes, they just state that changes occur, but without specifying the reason for change. Thus, the concept of ‘forces that drive organizational change’ (and their features) is not considered. Therefore, the objective of this PhD thesis is to translate the knowledge of the forces that drive organizational change available in human organizations to MASbased organizations. These forces will be formally expressed with the factors that help to detect them. The solutions to be taken when a force is detected will also be presented. To correctly perform this task, a formalization for virtual organizations is designed, named Virtual Organization Formalization (VOF). Moreover, the Artifacts for Organizational Mechanisms are proposed, which are a tool to help in the representation of organizational knowledge and in the modeling of the environment of the organization. This tool is based on the Agents & Artifacts (A&A) framework. / Esparcia García, S. (2015). Integrating driving forces into the development of Adaptive Virtual Organizations [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/48538

We propose a Model-Based Deep Reinforcement Learning (MBDRL) framework for collaborative paylaod transportation using Unmanned Aerial Vehicles (UAVs) in Search and Rescue (SAR) missions, enabling heavier payload conveyance while maintaining vehicle agility. Our approach extends the single-drone application to a novel multi-drone one, using the Probabilistic Ensembles with Trajectory Sampling (PETS) algorithm to model the unknown stochastic system dynamics and uncertainty. We use the Multi-Agent Reinforcement Learning (MARL) framework via a centralized controller in a leader-follower configuration. The agents utilize the approximated transition function in a Model Predictive Controller (MPC) configured to maximize the reward function for waypoint navigation, while a position-based formation controller ensures stable flights of these physically linked UAVs. We also developed an Unreal Engine (UE) simulation connected to an offboard planner and controller via a Robot Operating System (ROS) framework that is transferable to real robots. This work achieves stable waypoint navigation in a stochastic environment with a sample efficiency following that seen in single UAV work. This work has been funded by the National Science Foundation (NSF) under Award No. 2046770. / Master of Science / We apply the Model-Based Deep Reinforcement Learning (MBDRL) framework to the novel application of a UAV team transporting a suspended payload during Search and Rescue missions. Collaborating UAVs can transport heavier payloads while staying agile, reducing the need for human involvement. We use the Probabilistic Ensemble with Trajectory Sampling (PETS) algorithm to model uncertainties and build on the previously used single UAVpayload system. By utilizing the Multi-Agent Reinforcement Learning (MARL) framework via a centralized controller, our UAVs learn to transport the payload to a desired position while maintaining stable flight through effective cooperation. We also develop a simulation in Unreal Engine (UE) connected to a controller using a Robot Operating System (ROS) architecture, which can be transferred to real robots. Our method achieves stable navigation in unpredictable environments while maintaining the sample efficiency observed in single UAV scenarios.

This thesis explores the possibility of combining a noise-based terrain generator with a multi-agent system-based citygenerator for use in overworld map generation for games.The resulting artifact is a generator capable of creating awide variety of landscapes and populating them with citiesthat take advantage of available space and resources. Thegenerated noise is used as a height map which is colorizedto form the basis of the terrain. City agents traverse the space, looking for an optimal place to settle within their view. Once they find their final position, each city creates a number of building agents based on the nearby resources. The city agents take each other and the terrain into account when deciding where the buildings shall be placed. Then, road agents connectfirst the houses, then the cities together. Lastly, points ofinterests such as dungeons, bandit camps and wizard towers are placed, their final positions determined in a similar fashion to the city agents. The analysis of the results indicate that the generator is capable of producing novel terrain, and cities that adapt somewhat well to their surroundings, but offer little variety in existing features from map to map. In addition, more control is necessary to implement the artifact into adevelopment workflow. Respondents rated the maps produced by the generator lower than a map from an existing game ,converted into a similar format, and on a similar level as another map, also from an existing game.