Kyoto University (京都大学) / 0048 / 新制・課程博士 / 博士(工学) / 甲第15354号 / 工博第3233号 / 新制||工||1486(附属図書館) / 27832 / 京都大学大学院工学研究科都市社会工学専攻 / (主査)教授 岡田 憲夫, 教授 戸田 圭一, 教授 多々納 裕一 / 学位規則第4条第1項該当

The concept of consensus filters for sensor fusion is not an entirely new proposition but one with an internally implemented Bayesian fusion is. This work documents a novel state update algorithm for sensor fusion which works using the principle of Bayesian fusion of data with variance implemented on a single integrator consensus algorithm. Comparative demonstrations of how consensus over a pinning network is reached are presented along with a weighted Bayesian Luenberger type observer and a ’Consensus on estimates’ algorithm. This type of a filter is something that is novel and has not been encountered in previous literature related to this topic to the best of our knowledge. In this work, we also extend the proof for a distributed Luenberger type observer design to include the case where the network being considered is a strongly connected digraph.

In near future, robots are envisioned to work alongside humans in professional anddomestic environments without significant restructuring of workspace. Roboticsystems in such setups must be adept at observation, analysis and rational de-cision making. To coexist in an environment, humans and robots will need tointeract and cooperate for multiple tasks. A fundamental such task is the manip-ulation of large objects in work environments which requires cooperation betweenmultiple manipulating agents for load sharing. Collaborative manipulation hasbeen studied in the literature with the focus on multi-agent planning and controlstrategies. However, for a collaborative manipulation task, grasp planning alsoplays a pivotal role in cooperation and task completion.In this work, a novel approach is proposed for collaborative grasping and manipu-lation of large unknown objects. The manipulation task was defined as a sequenceof poses and expected external wrench acting on the target object. In a two-agentmanipulation task, the proposed approach selects a grasp for the second agentafter observing the grasp location of the first agent. The solution is computed ina way that it minimizes the grasp wrenches by load sharing between both agents.To verify the proposed methodology, an online system for human-robot manipu-lation of unknown objects was developed. The system utilized depth informationfrom a fixed Kinect sensor for perception and decision making for a human-robotcollaborative lift-up. Experiments with multiple objects substantiated that theproposed method results in an optimal load sharing despite limited informationand partial observability.

This dissertation aims at solving the consensus control problem of multi-agent systems with Lipschitz nonlinearity. This depends on the design of the controller that enables each agent or subsystem in multi-agent systems with Lipschitz nonlinearity to reach consensus; using the understanding of the agents' connection network from the knowledge of graph theory as well as the control system design strategy. The objective is achieved by designing a type of distributed control, namely the consensus control, which manipulates the relative information of each agent in a multi-agent systems in order to arrive at a single solution. In addition, containment control is also developed to solve containment problem. It is an extension of consensus control via leader-follower configuration, aimed at having each agent contained by multiple leaders in a multi-agent systems with Lipschitz nonlinearity. Four types of controllers are proposed - state-feedback consensus controller, observer-based consensus controller, state-feedback containment controller and observer-based containment controller; each provides the stability conditions based on Lyapunov stability analysis in time domain which enabled each agent or subsystem to reach consensus. The observer-based controllers are designed based on the consensus observer that is related to Luenberger observer. Linear Matrix Inequality (LMI) and Algebraic Riccati Equation (ARE) are utilized to obtain the solutions for the stability conditions. The simulation results of the proposed controllers and observers have been carried out to prove their theoretical validity. Several practical examples of flexible robot arm simulations are included to further validate the theoretical aspects of the thesis.

Nous élaborons une approche multi-agents pour la résolution des problèmes d’optimisation combinatoire nommée MAOM-COP. Elle combine des métaheuristiques, les systèmes multi-agents et l’apprentissage par renforcement. Les heuristiques manquent d’une vue d’ensemble sur l’évolution de la recherche. Notre objectif consiste à utiliser les systèmes multi-agents pour créer des méthodes de recherche coopératives. Ces méthodes explorent plusieurs métaheuristiques. MAOM-COP est composée de plusieurs agents qui sont l’agent décideur, les agents intensificateurs et les agents diversificateurs (agents croisement et agent perturbation). A l’aide de l’apprentissage, l’agent décideur décide dynamiquement quel agent à activer entre les agents intensificateurs et les agents croisement. Si les agents intensificateurs sont activés, ils appliquent des algorithmes de recherche locale. Durant leurs recherches, ils peuvent s’échanger des informations, comme ils peuvent déclencher l’agent perturbation. Si les agents croisement sont activés, ils exécutent des opérateurs de recombinaison. Nous avons appliqué MAOM-COP sur les problèmes suivants : l’affectation quadratique, la coloration des graphes, la détermination des gagnants et le sac à dos multidimensionnel. MAOM-COP possède des performances compétitives par rapport aux algorithmes de l’état de l’art. / We elaborate a multi-agent based optimization method for combinatorial optimization problems named MAOM-COP. It combines metaheuristics, multiagent systems and reinforcement learning. Although the existing heuristics contain several techniques to escape local optimum, they do not have an entire vision of the evolution of optimization search. Our main objective consists in using the multi-agent system to create intelligent cooperative methods of search. These methods explore several existing metaheuristics. MAOMCOP is composed of the following agents: the decisionmaker agent, the intensification agents and the diversification agents which are composed of the perturbation agent and the crossover agents. Based on learning techniques, the decision-maker agent decides dynamically which agent to activate between intensification agents and crossover agents. If the intensifications agents are activated, they apply local search algorithms. During their searches, they can exchange information, as they can trigger the perturbation agent. If the crossover agents are activated, they perform recombination operations. We applied MAOMCOP to the following problems: quadratic assignment, graph coloring, winner determination and multidimensional knapsack. MAOM-COP shows competitive performances compared with the approaches of the literature

Les cartes représentent l'information géographique d'une zone donnée de manière d'autant plus simplifiée que l'échelle de la carte est petite. Le procédé de simplification, appelé généralisation cartographique, est soumis au respect de contraintes de lisibilité, d'adéquation de la représentation avec le niveau d'abstraction souhaité et de cohérence avec la réalité. La volonté d'automatiser le processus de création de cartes à partir de bases de données géographiques, a conduit à la création d'algorithmes permettant d'effectuer cette simplification objet par objet. Néanmoins, les choix des algorithmes, tout comme leur paramétrage, sont autant influencés par l'objet sur lequel ils s'appliquent que par les autres objets en relation (e.g. bâtiment à proximité d'un autre, route parallèle à un alignement de bâtiments). Ce constat a motivé l'utilisation de modèles multi-agents pour la généralisation automatisée de cartes. Le principe de ces modèles multi-agents repose sur la modélisation des objets (e.g. bâtiment, tronçon de route, îlot urbain) sous forme d'agents qui cherchent à se généraliser de façon à satisfaire leurs contraintes. Plusieurs modèles multi-agents ont été proposés, chacun ayant une approche différente des interactions entre niveaux. Ici, nous entendons par niveau, par exemple, la distinction entre les agents individuels comme un bâtiment, des agents représentant un groupe d’autres agents, comme un îlot urbain composé des routes l’entourant et des bâtiments inclus dans l’îlot.Nous étudions l'unification de ces modèles en nous appuyant sur le paradigme multi-niveaux PADAWAN, afin de faciliter les interactions entre agents de niveaux différents. Nous proposons ainsi le modèle DIOGEN, adaptant les principes d’interaction entre agents de niveaux différents à la généralisation cartographique guidée par des contraintes, ce qui a permis d’unifier les précédents modèles AGENT, CartACom et GAEL, tout en disposant de nouvelles capacités prometteuses.Nous avons évalué notre proposition sur un ensemble de cas d’étude. Parmi ces cas, nous nous sommes penchés sur la généralisation de carte de randonnée, où les itinéraires sont symbolisés individuellement avec des symboles différents, à la manière des plans de bus. La présence de plusieurs symboles d’itinéraires sur une même route support amène des problèmes de généralisation particuliers, comme le choix du positionnement des itinéraires de part et d’autre de la route, ou les implications pour les autres objets de la carte (e.g. points d’intérêts, bâtiments) se retrouvant sous le symbole de l’itinéraire, problèmes que nous essayons de résoudre en nous appuyant sur notre proposition de représentation formelle multi-niveaux.Ce travail nous a ensuite conduit à identifier des comportements multi-niveaux récurrents. Nous les avons exprimés de façon générique sous forme de patterns d’analyse, affranchies des spécificités de la généralisation cartographique, et de la résolution de problèmes contraints / Maps show geographic information of a given area in a simplified way, particularly when the scale is small. The simplification process, called cartographic generalisation, is submitted to several constraints : legibility, adequation to the abstraction level, and consistency with reality. The will to automate the maps creation process from geographical databases led to the creation of algorithms allowing the simplification object by object. However the choice of the algorithms, as their settings, are influenced by the object on which it is applied, and by the other objects in relation with this object (e.g. a building close to another one, a road parallel to a buildings alignment). This motivates the use of multi-agents models for automated map generalisation. Several multi-agent models were proposed, each of them having a different approach to manage multi-levels relations. Here, what we call a level is, for instance, the distinction between individual agents, like a building, and agents representing a group of other agents, like a urban block composed by the surrounding roads and buildings inside.We study the unification of existing models, using the multi-level paradigm PADAWAN, in order to simplify interactions between agents in different levels. We propose the DIOGEN model, in which the principle of interactions between agents of different levels is adapted to cartographic generalisation guided by constraints, those allowing to unify the existing models AGENT, CartACom and GAEL, and giving promising features.We evaluate our proposal on different case studies. Among them, we study the generalisation of trekking maps, where the routes are symbolized individually by a different couloured line symbols, like on bus maps. The presence of several route symbols on a same road leads to specific generalisation issues, like the choice of the side of each route symbol position, or the implications for the other objects on the map (e.g. points of interest, buildings) under the route symbol – issues tackled using our proposal of formal multi-levels representation.This work leads us to the identification of recurrent behaviours. We express them as analysis patterns, in a way that is independent from cartographic generalisation and constraint solving problems

Multi-agent multi-team systems are commonly seen in environments where hierarchical layers of goals are at play. For example, theater-wide combat scenarios where multiple levels of command and control are required for proper execution of goals from the general to the foot soldier. Similar structures can be seen in game environments, where agents work together as teams to compete with other teams. The different agents within the same team must, while maintaining their own ‘personality’, work together and coordinate with each other to achieve a common team goal. This research develops strategy-based multi-agent multi-team systems, where strategy is framed as an instrument at the team level to coordinate the multiple agents of a team in a cohesive way. A formal specification of strategy and strategy-based multi-agent multi-team systems is provided. A framework is developed called SiMAMT (strategy- based multi-agent multi-team systems). The different components of the framework, including strategy simulation, strategy inference, strategy evaluation, and strategy selection are described. A graph-matching approximation algorithm is also developed to support effective and efficient strategy inference. Examples and experimental results are given throughout to illustrate the proposed framework, including each of its composite elements, and its overall efficacy. This research make several contributions to the field of multi-agent multi-team systems: a specification for strategy and strategy-based systems, and a framework for implementing them in real-world, interactive-time scenarios; a robust simulation space for such complex and intricate interaction; an approximation algorithm that allows for strategy inference within these systems in interactive-time; experimental results that verify the various sub-elements along with a full-scale integration experiment showing the efficacy of the proposed framework.

Cooperative control of multi-agent systems (MASs) has been intensively investigated in the past decade. The task is always complicated for an individual agent, but can be achieved by collectively operating a group of agents in a reliable, economic and efficient way. Although a lot of efforts are being spent on improving MAS performances, much progress has yet to be developed on different aspects. This thesis aims to solve problems in the consensus control of multiple quadrotors and/or mobile robots considering irregular sampling controls, heterogeneous agent dynamics and the presence of model uncertainties and disturbances. The thesis proceeds with Chapter 1 by providing the literature review of the state-of-the-art development in the consensus control of MASs. Chapter 2 introduces experimental setups of the laboratory involving two-wheeled mobile robots (2WMRs), quadrotors, positioning systems and inter-vehicle communications. All of the developed theoretical results in Chapters 3-6 are experimentally verified on the platform. Then it is followed by two main parts: Irregular sampling consensus control methods (Chapter 3 and 4) and cooperative control of heterogeneous MASs (Chapter 5 and 6). Chapter 3 focuses on the non-uniform sampling consensus control for a group of 2WMRs, and Chapter 4 studies the event-based rendezvous control for a group of asynchronous robots with time-varying communication delays. Chapter 5 concentrates on cooperative control methods for a heterogeneous MAS consisting of quadrotors and 2WMRs. Chapter 6 focuses on the design of a quadrotor flight controller which is robust to various adverse factors such as model uncertainties and external disturbances. The developed controller is further applied to the consensus control of the heterogeneous MAS. Specifically, Chapter 3 studies synchronized and non-periodical sampling consensus control methods for a group of 2WMRs. The directed and switching communication topologies among the network are considered in the controller design. The 2WMR is an underactuated system, which implies that it can not generate independent x and y accelerations in the two-dimensional plane. The rendezvous control methods are proposed for 2WMRs. The algebraic graph theory and stochastic matrix analysis are employed to conduct the convergence analysis. Although the samplings in the work of Chapter 3 are aperiodic, one feature is that local clocks of agents are required to be synchronized. Challenges arise in the practical control of distributed MASs, especially in the scenario that the global clock is lacking. Moreover, frequent samplings can result in redundant information transmissions when the communication bandwidth is limited. To address these problems, Chapter 4 investigates an event-based rendezvous control method for a group of asynchronous MAS with time-varying communication delays. Integral-type triggering conditions for each robot are adopted to be checked periodically. If the triggering condition is satisfied at one checking instant, the agent samples and broadcasts the state to the neighbors with a bounded communication delay. Then an algorithm is provided for driving 2WMRs to asymptotically reach rendezvous. The convergence analysis is conducted through Lyapunov approaches. Most of the theoretical works on cooperative control are focused on controlling agents with identical dynamics. However, in certain realistic scenarios, some complex missions require the cooperation of different types of agent dynamics such as surveillance, search and rescue, etc. Tasks can be carried out with higher efficiency by employing both the autonomous ground vehicles and unmanned aerial vehicles. To achieve better performance for MASs, in Chapter 5, distributed cooperative control methods for a heterogeneous MAS consisting of quadrotors and 2WMRs are developed. Consensus conditions are provided, and the theoretical results are experimentally verified. Many existing quadrotor control methods need exact model parameters of the quadrotor. In reality, when a quadrotor is conducting some tasks with extra payloads or with unexpected damages to the model structure, errors in parameters could result in the failure of the flight. External disturbances also inevitably affect the flight performance. To move a step further towards practical applications, in Chapter 6, a robust quadrotor flight controller using Integral Sliding Mode Control (ISMC) technique is investigated. In experiments, an extra payload with the position and mass unknown, is attached to destroy the accuracy of the model and to add disturbances. The designed controller significantly rejects negative effects caused by the payload during the flight. This controller is also successfully applied to an MAS consisting of a quadrotor and 2WMRs. / Graduate

Energy is consumed every day at home as we perform simple tasks, such as watching television, washing dishes and heating/cooling home spaces during season of extreme weather conditions, using appliances, or turning on lights. Most often, the energy resources used in residential systems are obtained from natural gas, coal and oil. Moreover, climate change has increased awareness of a need for expendable, energy resources. As a result, carbon dioxide emissions are increasing and creating a negative effect on our environment and on our health. In fact, growing energy demands and limited natural resource might have negative impacts on our future. Therefore, saving energy is becoming an important issue in our society and it is receiving more attention from the research community. This thesis introduces a intelligent energy controller algorithm based on software agent approach that reduce the energy consumption at home for both heating and cooling spaces by considering the user’s occupancy, outdoor temperature and user’s preferences as input to the system. Thus the proposed approach takes into consideration the occupant’s preferred temperature, the occupied and unoccupied spaces, as well as the time spent in each area of the home. A Java based simulator has been implemented to simulate the algorithm for saving energy in heating and cooling systems. The results from the simulator are compared to the results of using HOT2000, which is Canada’s leading residential energy analysis and rating software developed by CanmetENERGY’s Housing, Buildings, Communities and Simulation (HBCS) group. We have calculated how much energy a home modelled will use under emulated conditions. The results showed that the implementation of the proposed energy controller algorithm can save up to 50% in energy consumption in homes dedicated to heating and cooling systems compared to the results obtained by using HOT2000.

In this thesis we explore the importance of consumer protection (concretely protection against misleading advertising and minimal quality standards of product) for providing consumers' utility. Our target is to discover whether these protections are necessary for providing consumers' utility or not, and whether are the consumers able to provide themselves equal or higher utility (then can mentioned protections do), scilicet through their mutual communication and buying behavior. Furthermore, we want to discover, if can happen a situation in which the mentioned protection decline the consumers' utility, in comparison with situation, when no protections exist. These conclusions are stated for concrete product (LCD TV) and concrete society (little community within the Czech Republic society), whereas we explore the influence of consumer protection both on utility of whole community and on utility of each social class within different proportions of these classes within society. For this research has been created an original multi-agent model, in which we simulate consumer behavior and companies behavior on market. The important parts of the model architecture are presented in this thesis: principles of social structure creation (based on the knowledge about social groups and different types of social ties), buyer behavior and principles of consumer communication through social network, and decision of companies about product characteristics. In conclusion we have interpreted the simulation results and have stated an importance of consumer protection in tested society buying tested product.