Um modelo cognitivo baseado na atenção para consciência em robôs inteligentes

Understanding consciousness is one of the most fascinating challenges of our time. From ancient civilizations to modern philosophers, questions have been asked on how one is conscious of his/her own existence and about the world that surrounds him/her. Although there is no precise definition for consciousness, there is an agreement that it is strongly related to human cognitive processes such as: thinking, reasoning, emotions, wishes. One of the key processes to the arising of the consciousness is the attention, a process capable of promoting a selection of a few stimuli from a huge amount of information that reaches us constantly. Machine consciousness is the field of the artificial intelligence that investigate the possibility of the production of conscious processes in artificial devices. This work presents a review about the theme of consciousness - in both natural and artificial aspects -, discussing this theme from the philosophical and computational perspectives, and investigates the feasibility of the adoption of an attentional schema as the base to the cognitive processing. A formal computational model is proposed for conscious agents that integrates: short and long term memories, reasoning, planning, emotion, decision making, learning, motivation and volition. Computer experiments in a mobile robotics domain under USARSim simulation environment, proposed by RoboCup, suggest that the agent can be able to use these elements to acquire experiences based on environment stimuli. The adoption of the cognitive architecture over the attentional model has potential to allow the emergence of behaviours usually associated to the consciousness in the simulated mobile robots. Further implementation under this model could potentially allow the agent to express sentience, selfawareness, self-consciousness, autonoetic consciousness, mineness and perspectivalness. By performing computation over an attentional space, the model also allows the ...

Reinforcement Lernen mit Regularisierungsnetzwerken

Die Arbeit behandelt das Problem der Skalierbarkeit von Reinforcement Lernen auf hochdimensionale und komplexe Aufgabenstellungen. Unter Reinforcement Lernen versteht man dabei eine auf approximativem Dynamischen Programmieren basierende Klasse von Lernverfahren, die speziell Anwendung in der Künstlichen Intelligenz findet und zur autonomen Steuerung simulierter Agenten oder realer Hardwareroboter in dynamischen und unwägbaren Umwelten genutzt werden kann. Dazu wird mittels Regression aus Stichproben eine Funktion bestimmt, die die Lösung einer "Optimalitätsgleichung" (Bellman) ist und aus der sich näherungsweise optimale Entscheidungen ableiten lassen. Eine große Hürde stellt dabei die Dimensionalität des Zustandsraums dar, die häufig hoch und daher traditionellen gitterbasierten Approximationsverfahren wenig zugänglich ist. Das Ziel dieser Arbeit ist es, Reinforcement Lernen durch nichtparametrisierte Funktionsapproximation (genauer, Regularisierungsnetze) auf -- im Prinzip beliebig -- hochdimensionale Probleme anwendbar zu machen. Regularisierungsnetze sind eine Verallgemeinerung von gewöhnlichen Basisfunktionsnetzen, die die gesuchte Lösung durch die Daten parametrisieren, wodurch die explizite Wahl von Knoten/Basisfunktionen entfällt und so bei hochdimensionalen Eingaben der "Fluch der Dimension" umgangen werden kann. Gleichzeitig sind Regularisierungsnetze aber auch lineare Approximatoren, die technisch einfach handhabbar sind und für die die bestehenden Konvergenzaussagen von Reinforcement Lernen Gültigkeit behalten (anders als etwa bei Feed-Forward Neuronalen Netzen). Allen diesen theoretischen Vorteilen gegenüber steht allerdings ein sehr praktisches Problem: der Rechenaufwand bei der Verwendung von Regularisierungsnetzen skaliert von Natur aus wie O(n**3), wobei n die Anzahl der Daten ist. Das ist besonders deswegen problematisch, weil bei Reinforcement Lernen der Lernprozeß online erfolgt -- die Stichproben werden von einem Agenten/Roboter erzeugt, während er mit der Umwelt interagiert. Anpassungen an der Lösung müssen daher sofort und mit wenig Rechenaufwand vorgenommen werden. Der Beitrag dieser Arbeit gliedert sich daher in zwei Teile: Im ersten Teil der Arbeit formulieren wir für Regularisierungsnetze einen effizienten Lernalgorithmus zum Lösen allgemeiner Regressionsaufgaben, der speziell auf die Anforderungen von Online-Lernen zugeschnitten ist. Unser Ansatz basiert auf der Vorgehensweise von Recursive Least-Squares, kann aber mit konstantem Zeitaufwand nicht nur neue Daten sondern auch neue Basisfunktionen in das bestehende Modell einfügen. Ermöglicht wird das durch die "Subset of Regressors" Approximation, wodurch der Kern durch eine stark reduzierte Auswahl von Trainingsdaten approximiert wird, und einer gierigen Auswahlwahlprozedur, die diese Basiselemente direkt aus dem Datenstrom zur Laufzeit selektiert. Im zweiten Teil übertragen wir diesen Algorithmus auf approximative Politik-Evaluation mittels Least-Squares basiertem Temporal-Difference Lernen, und integrieren diesen Baustein in ein Gesamtsystem zum autonomen Lernen von optimalem Verhalten. Insgesamt entwickeln wir ein in hohem Maße dateneffizientes Verfahren, das insbesondere für Lernprobleme aus der Robotik mit kontinuierlichen und hochdimensionalen Zustandsräumen sowie stochastischen Zustandsübergängen geeignet ist. Dabei sind wir nicht auf ein Modell der Umwelt angewiesen, arbeiten weitestgehend unabhängig von der Dimension des Zustandsraums, erzielen Konvergenz bereits mit relativ wenigen Agent-Umwelt Interaktionen, und können dank des effizienten Online-Algorithmus auch im Kontext zeitkritischer Echtzeitanwendungen operieren. Wir demonstrieren die Leistungsfähigkeit unseres Ansatzes anhand von zwei realistischen und komplexen Anwendungsbeispielen: dem Problem RoboCup-Keepaway, sowie der Steuerung eines (simulierten) Oktopus-Tentakels.

Extending the expressiveness of fuzzy logic languages

En el trabajo que aquí presentamos se incluye la base teórica (sintaxis y semántica) y una implementación de un framework para codificar el razonamiento de la representación difusa o borrosa del mundo (tal y como nosotros, seres humanos, entendemos éste). El interés en la realización de éste trabajo parte de dos fuentes: eliminar la complejidad existente cuando se realiza una implementación con un lenguaje de programación de los llamados de propósito general y proporcionar una herramienta lo suficientemente inteligente para dar respuestas de forma constructiva a consultas difusas o borrosas. El framework, RFuzzy, permite codificar reglas y consultas en una sintaxis muy cercana al lenguaje natural usado por los seres humanos para expresar sus pensamientos, pero es bastante más que eso. Permite representar conceptos muy interesantes, como fuzzificaciones (funciones usadas para convertir conceptos no difusos en difusos), valores por defecto (que se usan para devolver resultados un poco menos válidos que los que devolveríamos si tuviésemos la información necesaria para calcular los más válidos), similaridad entre atributos (característica que utilizamos para buscar aquellos individuos en la base de datos con una característica similar a la buscada), sinónimos o antónimos y, además, nos permite extender el numero de conectivas y modificadores (incluyendo modificadores de negación) que podemos usar en las reglas y consultas. La personalización de la definición de conceptos difusos (muy útil para lidiar con el carácter subjetivo de los conceptos borrosos, donde nos encontramos con que cualificar a alguien de “alto” depende de la altura de la persona que cualifica) es otra de las facilidades incluida. Además, RFuzzy implementa la semántica multi-adjunta. El interés en esta reside en que introduce la posibilidad de obtener la credibilidad de una regla a partir de un conjunto de datos y una regla dada y no solo el grado de satisfacción de una regla a partir de el universo modelado en nuestro programa. De esa forma podemos obtener automáticamente la credibilidad de una regla para una determinada situación. Aún cuando la contribución teórica de la tesis es interesante en si misma, especialmente la inclusión del modificador de negacion, sus multiples usos practicos lo son también. Entre los diferentes usos que se han dado al framework destacamos el reconocimiento de emociones, el control de robots, el control granular en computacion paralela/distribuída y las busquedas difusas o borrosas en bases de datos. ABSTRACT In this work we provide a theoretical basis (syntax and semantics) and a practical implementation of a framework for encoding the reasoning and the fuzzy representation of the world (as human beings understand it). The interest for this work comes from two sources: removing the existing complexity when doing it with a general purpose programming language (one developed without focusing in providing special constructions for representing fuzzy information) and providing a tool intelligent enough to answer, in a constructive way, expressive queries over conventional data. The framework, RFuzzy, allows to encode rules and queries in a syntax very close to the natural language used by human beings to express their thoughts, but it is more than that. It allows to encode very interesting concepts, as fuzzifications (functions to easily fuzzify crisp concepts), default values (used for providing results less adequate but still valid when the information needed to provide results is missing), similarity between attributes (used to search for individuals with a characteristic similar to the one we are looking for), synonyms or antonyms and it allows to extend the number of connectives and modifiers (even negation) we can use in the rules. The personalization of the definition of fuzzy concepts (very useful for dealing with the subjective character of fuzziness, in which a concept like tall depends on the height of the person performing the query) is another of the facilities included. Besides, RFuzzy implements the multi-adjoint semantics. The interest in them is that in addition to obtaining the grade of satisfaction of a consequent from a rule, its credibility and the grade of satisfaction of the antecedents we can determine from a set of data how much credibility we must assign to a rule to model the behaviour of the set of data. So, we can determine automatically the credibility of a rule for a particular situation. Although the theoretical contribution is interesting by itself, specially the inclusion of the negation modifier, the practical usage of it is equally important. Between the different uses given to the framework we highlight emotion recognition, robocup control, granularity control in parallel/distributed computing and flexible searches in databases.

Modeling and exploiting behavior patterns in dynamic environments

This paper presents a new approach to improving the effectiveness of autonomous systems that deal with dynamic environments. The basis of the approach is to find repeating patterns of behavior in the dynamic elements of the system, and then to use predictions of the repeating elements to better plan goal directed behavior. It is a layered approach involving classifying, modeling, predicting and exploiting. Classifying involves using observations to place the moving elements into previously defined classes. Modeling involves recording features of the behavior on a coarse grained grid. Exploitation is achieved by integrating predictions from the model into the behavior selection module to improve the utility of the robot's actions. This is in contrast to typical approaches that use the model to select between different strategies or plays. Three methods of adaptation to the dynamic features of the environment are explored. The effectiveness of each method is determined using statistical tests over a number of repeated experiments. The work is presented in the context of predicting opponent behavior in the highly dynamic and multi-agent robot soccer domain (RoboCup)

UQ RoboRoos: Achieving power and agility in a small size robot

The UQ RoboRoos have been developed to participate in the RoboCup robot soccer small size league over several years. RoboCup 2001 saw a focus on the mechanical design of the RoboRoos, with the introduction of an omni-directional drive system and a high power kicker. The change in mechanical design had implications for the rest of the system particularly navigation and multi-robot planning. In addition, the overhead vision system was upgraded to improve reliability and robustness.

Classifying an opponent's behaviour in robot soccer

This paper illustrates the prediction of opponent behaviour in a competitive, highly dynamic, multi-agent and partially observableenvironment, namely RoboCup small size league robot soccer. The performance is illustrated in the context of the highly successful robot soccer team, the RoboRoos. The project is broken into three tasks; classification of behaviours, modelling and prediction of behaviours and integration of the predictions into the existing planning system. A probabilistic approach is taken to dealing with the uncertainty in the observations and with representing the uncertainty in the prediction of the behaviours. Results are shown for a classification system using a Naïve Bayesian Network that determines the opponent’s current behaviour. These results are compared to an expert designed fuzzy behaviour classification system. The paper illustrates how the modelling system will use the information from behaviour classification to produce probability distributions that model the manner with which the opponents perform their behaviours. These probability distributions are show to match well with the existing multi-agent planning system (MAPS) that forms the core of the RoboRoos system.

Decision Making for Teams of Mobile Robots

In the past years, we could observe a significant amount of new robotic systems in science, industry, and everyday life. To reduce the complexity of these systems, the industry constructs robots that are designated for the execution of a specific task such as vacuum cleaning, autonomous driving, observation, or transportation operations. As a result, such robotic systems need to combine their capabilities to accomplish complex tasks that exceed the abilities of individual robots. However, to achieve emergent cooperative behavior, multi-robot systems require a decision process that copes with the communication challenges of the application domain. This work investigates a distributed multi-robot decision process, which addresses unreliable and transient communication. This process composed by five steps, which we embedded into the ALICA multi-agent coordination language guided by the PROViDE negotiation middleware. The first step encompasses the specification of the decision problem, which is an integral part of the ALICA implementation. In our decision process, we describe multi-robot problems by continuous nonlinear constraint satisfaction problems. The second step addresses the calculation of solution proposals for this problem specification. Here, we propose an efficient solution algorithm that integrates incomplete local search and interval propagation techniques into a satisfiability solver, which forms a satisfiability modulo theories (SMT) solver. In the third decision step, the PROViDE middleware replicates the solution proposals among the robots. This replication process is parameterized with a distribution method, which determines the consistency properties of the proposals. In a fourth step, we investigate the conflict resolution. Therefore, an acceptance method ensures that each robot supports one of the replicated proposals. As we integrated the conflict resolution into the replication process, a sound selection of the distribution and acceptance methods leads to an eventual convergence of the robot proposals. In order to avoid the execution of conflicting proposals, the last step comprises a decision method, which selects a proposal for implementation in case the conflict resolution fails. The evaluation of our work shows that the usage of incomplete solution techniques of the constraint satisfaction solver outperforms the runtime of other state-of-the-art approaches for many typical robotic problems. We further show by experimental setups and practical application in the RoboCup environment that our decision process is suitable for making quick decisions in the presence of packet loss and delay. Moreover, PROViDE requires less memory and bandwidth compared to other state-of-the-art middleware approaches.

Dynamic set pieces for CAMBADA team

Robotics is an emergent branch of engineering that involves the conception, manufacture, and control of robots. It is a multidisciplinary field that combines electronics, design, computer science, artificial intelligence, mechanics and nanotechnology. Its evolution results in machines that are able to perform tasks with some level of complexity. Multi-agent systems is a researching topic within robotics, thus they allow the solving of higher complexity problems, through the execution of simple routines. Robotic soccer allows the study and development of robotics and multiagent systems, as the agents have to work together as a team, having in consideration most problems found in our quotidian, as for example adaptation to a highly dynamic environment as it is the one of a soccer game. CAMBADA is the robotic soccer team belonging to the group of research IRIS from IEETA, composed by teachers, researchers and students of the University of Aveiro, which annually has as main objective the participation in the RoboCup, in the Middle Size League. The purpose of this work is to improve the coordination in set pieces situations. This thesis introduces a new behavior and the adaptation of the already existing ones in the offensive situation, as well as the proposal of a new positioning method in defensive situations. The developed work was incorporated within the competition software of the robots. Which allows the presentation, in this dissertation, of the experimental results obtained, through simulation software as well as through the physical robots on the laboratory.

Goal detection using laser range finder for goalkeeper and striker from CAMBADA team

For a robot be autonomous and mobile, it requires being attached with a set of sensors that helps it to have a better perception of the surrounding world, to manage to localize itself and the surrounding objects. CAMBADA is the robotic soccer team of the IRIS research group, from IEETA, University of Aveiro, that competes in the Middle-Size League of RoboCup. In competition, in order to win, the main objective of the game it's to score more goals than the conceded, so not conceding goals, and score as much as possible it's desirable, thus, this thesis focus on adapt an agent with a better localization capacity in defensive and offensive moments. It was introduced a laser range finder to the CAMBADA robots, making them capable of detecting their own and the opponent goal, and to detect the opponents in specific game situations. With the new information and adapting the Goalie and Penalty behaviors, the CAMBADA goalkeeper is now able to detect and track its own goal and the CAMBADA striker has a better performance in a penalty situation. The developed work was incorporated within the competition software of the robots, which allows the presentation, in this thesis, of the experimental results obtained with physical robots on the laboratory field.

Arquitectura de Comportamientos Reactivos para Agentes Robóticos basada en CBR

En los últimos tiempos se ha demostrado la importancia del aprendizaje en la Inteligencia humana, tanto en su vertiente de aprendizaje por observación como a través de la experiencia, como medio de identificar situaciones y predecir acciones o respuestas a partir de la información adquirida. Dado este esquema general de la Inteligencia Humana, parece razonable imitar su estructura y características en un intento por diseñar una arquitectura general de inteligencia aplicada a la Robótica. En este trabajo, inspirados por las teorías de Hawkins en su obra On Intelligence, hemos propuesto una arquitectura jerárquica de inteligencia en el que los diversos módulos se implementan a partir de Razonamiento basado en Casos ¿Case Based Reasoning (CBR)¿, una herramienta de IA especialmente apta para la adquisición de conocimiento a través del aprendizaje y para la predicción basada en similitud de información. Dentro de esta arquitectura la presente tesis se centra en las capas inferiores, las de tipo reactivo, expresadas en forma de comportamientos básicos, que implementan conductas sencillas pero indispensables para el funcionamiento de un robot. Estos comportamientos han sido tradicionalmente diseñados de forma algorítmica, con la dificultad que esto entraña en muchos casos por el desconocimiento de sus aspectos intrínsecos. Además, carecen de la capacidad de adaptarse ante nuevas situaciones no previstas y adquirir nuevos conocimientos a través del funcionamiento del robot, algo indispensable si se pretende que éste se desenvuelva en ambientes dinámicos y no estructurados. El trabajo de esta tesis considera la implementación de comportamientos reactivos con capacidad de aprendizaje, como forma de superar los inconvenientes anteriormente mencionados consiguiendo al mismo tiempo una mejor integración en la arquitectura general de Inteligencia considerada, en la cual el aprendizaje ocupa el papel principal. Así, se proponen y analizan diversas alternativas de diseño de comportamientos reactivos, construidos a través de sistemas CBR con capacidad de aprendizaje. En particular se estudia i) la problemática de selección, organización, y representación de la información como recipiente del conocimiento de los comportamientos;ii) los problemas asociados a la escalabilidad de esta información; iii) los aspectos que acompañan al proceso de predicción mediante la recuperación de la respuesta de experiencias previas similares a la presentada; iv) la identificación de la respuesta no solo con la acción a tomar por parte del comportamiento sino con un concepto que represente la situación presentada; y v) la adaptación y evaluación de la respuesta para incorporar nuevas situaciones como nuevo conocimiento del sistema. También se analiza la organización de comportamientos básicos que permite obtener, a través de sus interacciones, comportamientos emergentes de nivel superior aún dentro de un alcance reactivo. Todo ello se prueba con un robot real y con un simulador, en una variante de un escenario de aplicación clásico en Robótica, como es la competición Robocup. La elaboración de esta tesis ha supuesto, además de los aspectos puramente investigadores, un esfuerzo adicional en el desarrollo de las herramientas y metodología de pruebas necesarias para su realización. En este sentido, se ha programado un primer prototipo de marco de implementación de comportamientos reactivos con aprendizaje, basados en CBR, para la plataforma de desarrollo robótico Tekkotsu.