The Intelligent Control Paradigm for software architecture is the result of this work. The Viable Software Approach is developed as an instance of the paradigm. The approach uses the Viable System Model as the basis for software system architecture. The result is a model-based architecture and approach for developing software systems by piecemeal adaptation with the goal that they become adaptive systems at runtime. Software built in this manner is called Viable Software. Viable Software represents a unifying class of self-controlling software that is an “intelligent” control system. Cybernetics, Control Theory, and Complexity Theory are the background for this work, and aspects relevant to this work are presented. These results are related to software architecture and software engineering. Rationale for the selection of the Viable System Model as a basis for software systems is given. The Viable System Model is described. The model is restated as an Alexanderian “pattern language” to make it more accessible to software engineering. A Viable Software Approach is proposed and expressed in the form of a Product Line Architecture that arranges the Viable System Model, the Viable Software Architecture, a Viable Component Framework, and a Component Transfer Protocol into a system for generative programming. An important result is the formalisation of the pattern of the Viable System into the interface specifications of the Viable Component. Three case studies illustrate the approach. The first is an analysis and extension of the Groove collaboration system. This study shows how the approach is used to map an existing system into the Viable Software Architecture and add fuzzy-adaptive user interface controllers. The second study presents the design and detailed software construction of an adaptive camera controller as part of a smart environment. The final study shows how a Business-to-Business e-Commerce system can be evolved and an expert system-based controller developed to implement business contracts.

[Truncated abstract. Please see the pdf version for the complete text. Also, formulae and special characters can only be approximated here. Please see the pdf version of this abstract for an accurate reproduction.] This thesis treats two aspects of intelligent control: The first part is about long-term optimization by approximating dynamic programming and in the second part a specific class of a fast neural network, related to support vector machines (SVMs), is considered. The first part relates to approximate dynamic programming, especially in the framework of adaptive critic designs (ACDs). Dynamic programming can be used to find an optimal decision or control policy over a long-term period. However, in practice it is difficult, and often impossible, to calculate a dynamic programming solution, due to the 'curse of dimensionality'. The adaptive critic design framework addresses this issue and tries to find a good solution by approximating the dynamic programming process for a stationary environment. In an adaptive critic design there are three modules, the plant or environment to be controlled, a critic to estimate the long-term cost and an action or controller module to produce the decision or control strategy. Even though there have been many publications on the subject over the past two decades, there are some points that have had less attention. While most of the publications address the training of the critic, one of the points that has not received systematic attention is training of the action module.¹ Normally, training starts with an arbitrary, hopefully stable, decision policy and its long-term cost is then estimated by the critic. Often the critic is a neural network that has to be trained, using a temporal difference and Bellman's principle of optimality. Once the critic network has converged, a policy improvement step is carried out by gradient descent to adjust the parameters of the controller network. Then the critic is retrained again to give the new long-term cost estimate. However, it would be preferable to focus more on extremal policies earlier in the training. Therefore, the Calculus of Variations is investigated to discard the idea of using the Euler equations to train the actor. However, an adaptive critic formulation for a continuous plant with a short-term cost as an integral cost density is made and the chain rule is applied to calculate the total derivative of the short-term cost with respect to the actor weights. This is different from the discrete systems, usually used in adaptive critics, which are used in conjunction with total ordered derivatives. This idea is then extended to second order derivatives such that Newton's method can be applied to speed up convergence. Based on this, an almost concurrent actor and critic training was proposed. The equations are developed for any non-linear system and short-term cost density function and these were tested on a linear quadratic regulator (LQR) setup. With this approach the solution to the actor and critic weights can be achieved in only a few actor-critic training cycles. Some other, more minor issues, in the adaptive critic framework are investigated, such as the influence of the discounting factor in the Bellman equation on total ordered derivatives, the target interpretation in backpropagation through time as moving and fixed targets, the relation between simultaneous recurrent networks and dynamic programming is stated and a reinterpretation of the recurrent generalized multilayer perceptron (GMLP) as a recurrent generalized finite impulse MLP (GFIR-MLP) is made. Another subject in this area that is investigated, is that of a hybrid dynamical system, characterized as a continuous plant and a set of basic feedback controllers, which are used to control the plant by finding a switching sequence to select one basic controller at a time. The special but important case is considered when the plant is linear but with some uncertainty in the state space and in the observation vector, and a quadratic cost function. This is a form of robust control, where a dynamic programming solution has to be calculated. &sup1Werbos comments that most treatment of action nets or policies either assume enumerative maximization, which is good only for small problems, except for the games of Backgammon or Go [1], or, gradient-based training. The latter is prone to difficulties with local minima due to the non-convex nature of the cost-to-go function. With incremental methods, such as backpropagation through time, calculus of variations and model-predictive control, the dangers of non-convexity of the cost-to-go function with respect to the control is much less than the with respect to the critic parameters, when the sampling times are small. Therefore, getting the critic right has priority. But with larger sampling times, when the control represents a more complex plan, non-convexity becomes more serious.

Thesis (MScEng)--Stellenbosch University, 2012. / ENGLISH ABSTRACT: Simultaneous localization and mapping (SLAM) is vital for autonomous robot navigation. The robot must build a map of its environment while tracking its own motion through that map. Although many solutions to this intricate problem have been proposed, one of the most prominent issues that still needs to be resolved is to accurately measure and track landmarks over time. In this thesis we investigate the use of stereo vision for this purpose. In order to find landmarks in images we explore the use of two feature detectors: the scale-invariant feature transform (SIFT) and speeded-up robust features (SURF). Both these algorithms find salient points in images and calculate a descriptor for each point that is invariant to scale, rotation and illumination. By using the descriptors we match these image features between stereo images and use the geometry of the system to calculate a set of 3D landmark measurements. A Taylor approximation of this transformation is used to derive a Gaussian noise model for the measurements. The measured landmarks are matched to landmarks in a map to find correspondences. We find that this process often incorrectly matches ambiguous landmarks. To find these mismatches we develop a novel outlier detection scheme based on the random sample consensus (RANSAC) framework. We use a similarity transformation for the RANSAC model and derive a probabilistic consensus measure that takes the uncertainties of landmark locations into account. Through simulation and practical tests we find that this method is a significant improvement on the standard approach of using the fundamental matrix. With accurately identified landmarks we are able to perform SLAM. We investigate the use of three popular SLAM algorithms: EKF SLAM, FastSLAM and FastSLAM 2. EKF SLAM uses a Gaussian distribution to describe the systems states and linearizes the motion and measurement equations with Taylor approximations. The two FastSLAM algorithms are based on the Rao-Blackwellized particle filter that uses particles to describe the robot states, and EKFs to estimate the landmark states. FastSLAM 2 uses a refinement process to decrease the size of the proposal distribution and in doing so decreases the number of particles needed for accurate SLAM. We test the three SLAM algorithms extensively in a simulation environment and find that all three are capable of very accurate results under the right circumstances. EKF SLAM displays extreme sensitivity to landmark mismatches. FastSLAM, on the other hand, is considerably more robust against landmark mismatches but is unable to describe the six-dimensional state vector required for 3D SLAM. FastSLAM 2 offers a good compromise between efficiency and accuracy, and performs well overall. In order to evaluate the complete system we test it with real world data. We find that our outlier detection algorithm is very effective and greatly increases the accuracy of the SLAM systems. We compare results obtained by all three SLAM systems, with both feature detection algorithms, against DGPS ground truth data and achieve accuracies comparable to other state-of-the-art systems. From our results we conclude that stereo vision is viable as a sensor for SLAM. / AFRIKAANSE OPSOMMING: Gelyktydige lokalisering en kartering (simultaneous localization and mapping, SLAM) is ’n noodsaaklike proses in outomatiese robot-navigasie. Die robot moet ’n kaart bou van sy omgewing en tegelykertyd sy eie beweging deur die kaart bepaal. Alhoewel daar baie oplossings vir hierdie ingewikkelde probleem bestaan, moet een belangrike saak nog opgelos word, naamlik om landmerke met verloop van tyd akkuraat op te spoor en te meet. In hierdie tesis ondersoek ons die moontlikheid om stereo-visie vir hierdie doel te gebruik. Ons ondersoek die gebruik van twee beeldkenmerk-onttrekkers: scale-invariant feature transform (SIFT) en speeded-up robust features (SURF). Altwee algoritmes vind toepaslike punte in beelde en bereken ’n beskrywer vir elke punt wat onveranderlik is ten opsigte van skaal, rotasie en beligting. Deur die beskrywer te gebruik, kan ons ooreenstemmende beeldkenmerke soek en die geometrie van die stelsel gebruik om ’n stel driedimensionele landmerkmetings te bereken. Ons gebruik ’n Taylor- benadering van hierdie transformasie om ’n Gaussiese ruis-model vir die metings te herlei. Die gemete landmerke se beskrywers word dan vergelyk met dié van landmerke in ’n kaart om ooreenkomste te vind. Hierdie proses maak egter dikwels foute. Om die foutiewe ooreenkomste op te spoor het ons ’n nuwe uitskieterherkenningsalgoritme ontwikkel wat gebaseer is op die RANSAC-raamwerk. Ons gebruik ’n gelykvormigheidstransformasie vir die RANSAC-model en lei ’n konsensusmate af wat die onsekerhede van die ligging van landmerke in ag neem. Met simulasie en praktiese toetse stel ons vas dat die metode ’n beduidende verbetering op die standaardprosedure, waar die fundamentele matriks gebruik word, is. Met ons akkuraat geïdentifiseerde landmerke kan ons dan SLAM uitvoer. Ons ondersoek die gebruik van drie SLAM-algoritmes: EKF SLAM, FastSLAM en FastSLAM 2. EKF SLAM gebruik ’n Gaussiese verspreiding om die stelseltoestande te beskryf en Taylor-benaderings om die bewegings- en meetvergelykings te lineariseer. Die twee FastSLAM-algoritmes is gebaseer op die Rao-Blackwell partikelfilter wat partikels gebruik om robottoestande te beskryf en EKF’s om die landmerktoestande af te skat. FastSLAM 2 gebruik ’n verfyningsproses om die grootte van die voorstelverspreiding te verminder en dus die aantal partikels wat vir akkurate SLAM benodig word, te verminder. Ons toets die drie SLAM-algoritmes deeglik in ’n simulasie-omgewing en vind dat al drie onder die regte omstandighede akkurate resultate kan behaal. EKF SLAM is egter baie sensitief vir foutiewe landmerkooreenkomste. FastSLAM is meer bestand daarteen, maar kan nie die sesdimensionele verspreiding wat vir 3D SLAM vereis word, beskryf nie. FastSLAM 2 bied ’n goeie kompromie tussen effektiwiteit en akkuraatheid, en presteer oor die algemeen goed. Ons toets die hele stelsel met werklike data om dit te evalueer, en vind dat ons uitskieterherkenningsalgoritme baie effektief is en die akkuraatheid van die SLAM-stelsels beduidend verbeter. Ons vergelyk resultate van die drie SLAM-stelsels met onafhanklike DGPS-data, wat as korrek beskou kan word, en behaal akkuraatheid wat vergelykbaar is met ander toonaangewende stelsels. Ons resultate lei tot die gevolgtrekking dat stereo-visie ’n lewensvatbare sensor vir SLAM is.

[pt] Um dos maiores desafios da produção de petróleo offshore é garantir um escoamento regular do reservatório até a unidade de processamento da produção. Entre os fenômenos que podem dificultar esta tarefa está o estabelecimento de um escoamento em golfadas no riser de produção, caracterizado por oscilações periódicas e de grande amplitude nas vazões do sistema, que diminuem a eficiência da planta de separação, sujeitam equipamentos a esforços cíclicos e causam graves descontroles de processo. Por esta razão, foram desenvolvidas nas últimas décadas diversas soluções que visam a evitar a formação das golfadas, ou ao menos, atenuar suas consequências. Entre as soluções mais promissoras estão os controladores que evitam a formação das golfadas através da manipulação ativa da válvula choke e os controladores que amortecem as golfadas nos vasos separadores da planta de processo. Neste trabalho, estas estratégias de controle são revisitadas sob a óptica do controle inteligente, possibilitando a obtenção de resultados fora do alcance dos controladores lineares. Além de um modelo computacional que descreve um sistema de produção do poço ao vaso separador, também foram desenvolvidos neste trabalho dois controladores inteligentes. O primeiro deles é um controlador anti-golfada de arquitetura híbrida Fuzzy-PID, que é capaz de suprimir as golfadas mesmo em sistemas desprovidos de medições submarinas e com válvula choke lenta. O segundo é um controlador amortecedor de golfadas fuzzy, otimizado por um algoritmo genético, com alta capacidade de atenuação das golfadas e capaz de manter o nível dentro de uma faixa considerada segura. Ambos os controladores são testados em diversos cenários e têm seus resultados comparados aos obtidos por controladores lineares. / [en] One of the major challenges of offshore oil production is to ensure a regular flow from the reservoir to the production processing unit. Among the phenomena that may hamper this task is the establishment of slug flow in the production riser, characterized by periodic oscillations of large amplitude in the system flow rates, which reduce the efficiency of the separation plant, subject equipment to cyclical fatigue and cause serious process instabilities. For this reason, several solutions that aim at avoiding the formation of slugs or, at least, mitigating their consequences have been developed in the last decades. Among the most promising solutions are controllers that prevent the formation of slugs by actively manipulating the choke valve and controllers that dampen the slugs in the vessels of the separation plant. In this work, these control strategies are revisited from perspective of intelligent control, allowing the obtainment of results beyond the reach of linear controllers. In addition to a computational model that comprises a production system from the well to the separation vessel, two intelligent controllers were developed in this work. One of them is a hybrid Fuzzy-PID anti-slug controller that is capable of suppressing slugs even in systems without submarine measurements available and with slow choke valve. The other one is a fuzzy slug damping controller, optimized by a genetic algorithm, with high slug attenuation capacity and able to maintain the level within a specified range. Both controllers are tested in several scenarios and have their results compared to those obtained by linear controllers.

Made available in DSpace on 2014-12-17T14:55:03Z (GMT). No. of bitstreams: 1 OscarGF.pdf: 1901439 bytes, checksum: f8f1a37dca7a69d726f7a9453cbf0a98 (MD5) Previous issue date: 2004-03-05 / This work develops a robustness analysis with respect to the modeling errors, being applied to the strategies of indirect control using Artificial Neural Networks - ANN s, belong to the multilayer feedforward perceptron class with on-line training based on gradient method (backpropagation). The presented schemes are called Indirect Hybrid Control and Indirect Neural Control. They are presented two Robustness Theorems, being one for each proposed indirect control scheme, which allow the computation of the maximum steady-state control error that will occur due to the modeling error what is caused by the neural identifier, either for the closed loop configuration having a conventional controller - Indirect Hybrid Control, or for the closed loop configuration having a neural controller - Indirect Neural Control. Considering that the robustness analysis is restrict only to the steady-state plant behavior, this work also includes a stability analysis transcription that is suitable for multilayer perceptron class of ANN s trained with backpropagation algorithm, to assure the convergence and stability of the used neural systems. By other side, the boundness of the initial transient behavior is assured by the assumption that the plant is BIBO (Bounded Input, Bounded Output) stable. The Robustness Theorems were tested on the proposed indirect control strategies, while applied to regulation control of simulated examples using nonlinear plants, and its results are presented / Este trabalho utiliza as Redes Neurais Multicamadas - RNM s, totalmente com treinamento em tempo real (on-line), no desenvolvimento de duas estrat?gias de controle indireto. Os esquemas propostos denominam-se Controle H?brido Indireto e Controle Neural Indireto. Todo o treinamento dos neurodispositivos - o identificador da planta e o controlador, quando presentes na malha de controle indireto, ? realizado com um m?nimo de atraso computacional, de modo a contemplar o controle de plantas com pequenos per?odos de amostragem. S?o apresentados Teoremas de Estabilidade para garantia da converg?ncia dos dispositivos neurais, assim como foram feitas considera??es para adequar o m?todo de acelera??o da converg?ncia h-adaptativo utilizado ?s condi??es de estabilidade. Para cada esquema de controle indireto foi desenvolvido um teorema que permite calcular o m?ximo erro permanente (steady-state error) que poder? ocorrer em fun??o da toler?ncia previamente especificada para converg?ncia dos dispositivos neurais usados na malha de controle, desde que a estabilidade seja garantida. Estes teoremas foram denominados de Teoremas da Robustez e constituem a principal contribui??o deste trabalho. As condi??es de estabilidade e robustez foram testadas para as estrat?gias de Controle H?brido Indireto e de Controle Neural Indireto, sendo apresentados os resultados obtidos na simula??o computacional do controle de regula??o de plantas n?o-lineares, BIBO (Bounded Input, Bounded Output) est?veis

Made available in DSpace on 2014-12-17T14:55:57Z (GMT). No. of bitstreams: 1 PedroBL.pdf: 595050 bytes, checksum: f0ab25c9935d2925102dfa8c9a811a4b (MD5) Previous issue date: 2005-12-16 / This work presents an analysis of the control law based on an indirect hybrid scheme using neural network, initially proposed for O. Adetona, S. Sathanathan and L. H. Keel. Implementations of this control law, for a level plant of second order, was resulted an oscillatory behavior, even if the neural identifier has converged. Such results had motivated the investigation of the applicability of that law. Starting from that, had been made stability mathematical analysis and several implementations, with simulated plants and with real plants, for analyze the problem. The analysis has been showed the law was designed being despised some components of dynamic of the plant to be controlled. Thus, for plants that these components have a significant influence in its dynamic, the law tends to fail / Este trabalho apresenta uma an?lise da lei de controle baseada em um esquema h?brido indireto usando rede neural, proposto inicialmente por O. Adetona, S. Sathanathan e L. H. Keel. Implementa??es dessa lei de controle, para uma planta de n?vel de segunda ordem, resultaram em um comportamento oscilat?rio, mesmo com a converg?ncia do identificador neural. Tais resultados motivaram a investiga??o da aplicabilidade dessa lei. A partir disso, foram feitas an?lises matem?ticas de estabilidade e diversas implementa??es, com plantas simuladas e com plantas reais, com a finalidade de se analisar o problema. A an?lise mostrou que a lei foi desenvolvida desprezando-se certos componentes da din?mica da planta a ser controlada. Sendo assim, para plantas onde esses componentes t?m uma influ?ncia significativa em sua din?mica, a lei tende a falhar

Submitted by Haroudo Xavier Filho (haroudo.xavierfo@ufpe.br) on 2016-04-06T18:16:38Z No. of bitstreams: 2 license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5) SISTEMA DE PREDIÇÃO DE ESTADOS DE TRANSDUTORES PARA AMBIEN.pdf: 2277201 bytes, checksum: 566bf1bfadd889426473a7471d98ef39 (MD5) / Made available in DSpace on 2016-04-06T18:16:38Z (GMT). No. of bitstreams: 2 license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5) SISTEMA DE PREDIÇÃO DE ESTADOS DE TRANSDUTORES PARA AMBIEN.pdf: 2277201 bytes, checksum: 566bf1bfadd889426473a7471d98ef39 (MD5) Previous issue date: 2015-08-26 / CNPq / Nos Ambientes Inteligentes, os dispositivos colaboram entre si para auxiliar o usuário de forma não intrusiva. Uma forma de auxílio é antecipar as ações do usuário e realizá-las por ele ou facilitar a sua realização. Esse trabalho propõe um framework para a predição das ações do usuário pelo aprendizado do seu comportamento e hábitos enquanto ele interage com o Ambiente Inteligente. As ações do usuário é considerada como sendo a troca do valor de um transdutor (sensor ou atuador). A interação do usuário com o Ambiente Inteligente produz o contexto que é utilizado para a predição das ações. O preditor é um algoritmo de classificação supervisionada que aprende os padrões de comportamento do habitante do Ambiente Inteligente. Portanto, a solução proposta pode prover um serviço personalizado e adaptativo ao invés de um conjunto de regras predefinido por humanos. O preditor trabalha apenas com um transdutor alvo e para prever valores de mais transdutores, mais preditores devem ser treinados. A solução proposta é projetada para funcionar automaticamente sem a necessidade de interferência humana. Isso faz com que o habitante do Ambiente Inteligente sinta-se mais confortável já que sua privacidade estará protegida. Todas as informações para treinar o preditor podem ser obtidas diretamente dos transdutores do Ambiente Inteligente. Não existe a necessidade de anotação manual dos dados e nem dados extras como tipo do transdutor, localização do transdutor ou objeto ao qual o transdutor está acoplado. Isso aumenta a facilidade de instalação dos transdutores no Ambiente Inteligente. A saída do preditor pode tanto controlar diretamente um atuador ou ser enviada a um agente de software. Esse agente pode verificar condições de segurança ou requisitos de gerenciamento de energia antes de tomar a decisão. O foco desse trabalho é a geração de uma base de dados com os dados do contexto para o treinamento do preditor responsável por decidir se o transdutor alvo deverá ou não mudar seu valor. Vários parâmetros são considerados como o tamanho do período de treinamento, quantidade de ativações passadas que serão consideradas e quais são os transdutores mais relevantes para a predição. A solução proposta atinge uma melhora significativa para todos os transdutores estudados e a maioria das combinações de parâmetros da geração da base de dados possuem resultados melhores que o caso base. Além disso, os nossos resultados são superiores às outras soluções da literatura. / Smart environments possess devices that collaborate to help the user non-intrusively. One possible aid smart environment offer is to anticipate user’s tasks and perform them on his/her behalf or facilitate the action completion. In this work, we propose a framework that predicts user’s actions by learning his/her behavior when interacting with the smart environment. The user actions are considered as being the value change of a transducer (sensor or actuator). The user interaction with the smart environment produces the context used to predict the actions. The predictor is a supervised classification algorithm that learns the smart environment inhabitant behavior patterns. Therefore, the proposed solution can provide a personalized and adaptive service instead of a human predefined set of rules. The predictor works with only one transducer and to predict the values of several transducers, more predictors must be trained. The proposed solution is designed to work automatically without the need of human interference. That makes the smart environment inhabitant more comfortable since his/her privacy is protected. All the information needed to train the predictor can be obtained directly from the smart environment transducers. There is no need for manual data annotation or extra data such as transducer type, transducer location or which object the transducer is attached to. This facilitates the transducer installation in the smart environment. The predictor output can either control directly an actuator or be sent to an software agent. This software agent can check for security or energy constraints before making the decision. This work focus on prepare datasets and train a predictor that is responsible to decide whether a target transducer value should be changed or not. Several parameters are considered such as the training period size, amount of previous transducer activations considered and which are the most relevant transducers for the prediction. Our solution achieves a significant improvement for all target transducers studied and most combinations of parameters yields better results than the base case. Our results are superior to other solutions in the literature.

Orientador: Fernando Jose Von Zuben / Acompanha CD-ROM / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-13T02:48:36Z (GMT). No. of bitstreams: 1 Cazangi_RenatoReder_D.pdf: 8716830 bytes, checksum: 272657e08f1aeb5622ebeb4412c49048 (MD5) Previous issue date: 2008 / Resumo: Novas técnicas de navegação autônoma de robôs móveis visam suprir a crescente demanda pelo emprego de robôs em diversos setores da sociedade e junto a uma ampla gama de tarefas. Os desafios envolvidos no desenvolvimento do sistema que controla o robô permitem afirmar que a inteligência embarcada em robôs atuais ainda encontra-se em um nível incipiente e limitado. Neste trabalho, cinco frentes de pesquisa complementares são propostas visando estudar, teórica e praticamente, aspectos fundamentais de projeto e implementação de controladores autônomos inteligentes para robótica móvel. Metodologias de computação bio-inspirada e de robótica evolutiva provêem os conceitos e ferramentas que fundamentam as cinco propostas, as quais são validadas com base em sistemas de navegação concebidos e aplicados a problemas relevantes da área. Uma série de simulações computacionais em ambientes virtuais e experimentos com robôs reais é realizada, permitindo medir o alcance das contribuições e apontar as principais frentes de atuação que se abrem como perspectivas futuras da pesquisa. / Abstract: Novel techniques for autonomous robot navigation aim at fulfilling the growing demand for mobile robots in multiple segments of society and in a plethora of tasks. The challenges involved in developing the system which controls the robot allow to say that the intelligence embedded in the current robots is found to be still incipient and limited. In this work, five complementary research fronts are proposed intending to study, theoretical and practically, aspects which are fundamental to the design and implementation of intelligent autonomous controllers for mobile robotics. Bio-inspired computing and evolutionary robotics methodologies provide the concepts and tools underlying the five proposals, which are validated through navigation systems devised and applied to important problems. Numerous real robot experiments as well as computational simulations taking place in virtual environments are carried out, allowing for the evaluation of contributions and also the discussion of future possibilities. / Doutorado / Engenharia de Computação / Doutor em Engenharia Elétrica

Orientador: Celso Pascoli Bottura / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-09T03:04:34Z (GMT). No. of bitstreams: 1 Grinits_ErickVile_D.pdf: 5949131 bytes, checksum: 59874252a786b0b7540c35faeb70854f (MD5) Previous issue date: 2007 / Resumo: Esta tese apresenta contribuições para o controle inteligente de sistemas não lineares com incertezas paramétricas e funcionais. São desenvolvidas duas abordagens. Em primeiro lugar, é proposta uma modificação ao backstepping adaptativo clássico, fundamentada em Extensão do Princípio de Invariância de La Salle que contempla o caso em que a derivada da função de Lyapunov ao longo das soluções do sistema é permitida ser definida positiva em regiões limitadas do espaço de estados, que favorece o pleno uso de técnicas de otimização baseadas em computação evolutiva com vistas à melhoria de desempenho da solução em termos de redução expressiva de esforço de controle e conformação da resposta transitória conforme critérios de tempo de estabilização, magnitude de sobre-sinal, etc. Em segundo lugar, é proposta uma metodologia de controle adaptativo neural, estruturada na técnica de backstepping, aplicável a sistemas não lineares com incertezas com múltiplas entradas e múltiplas saídas compostos de subsistemas interconectados cujas matrizes de entrada apresentam funções não lineares e cujas interconexões entre subsistemas também apresentam não linearidades. A estratégia apresentada propõe-se a sanar várias das dificuldades encontradas em metodologias análogas, eliminando a questão de singularidade nas leis de controle, evitando o uso de funções de Lyapunov com integrais e suprimindo a necessidade de introdução de derivadas dos controladores virtuais nas redes neurais, o que se traduz em projetos de controle menos complexos, com menor custo computacional e com melhor desempenho na solução em comparação aos métodos similares existentes / Abstract: This thesis presents contributions to the intelligent control of nonlinear systems with parametric and functional uncertainties. Two approaches are advanced. Firstly, it is proposed a modification to the traditional adaptive backstepping grounded on an Extension to the La Salle¿s Invariance Principle that allows the Lyapunov function derivative along the systems solutions to be positive definite in limited regions of the state space, which favours the deployment of optimisation techniques based on evolutionary computation aiming at the improvement of the solution performance in terms of an expressive reduction of control effort and transient conformation according to criteria such as settling time, overshoot magnitude, etc. Secondly, it is proposed a neural adaptive control methodology, structured on the backstepping technique, applicable to nonlinear systems with uncertainties with multiple inputs and multiple outputs constituted of interconnected subsystems whose input matrices feature nonlinear functions and whose subsystems interconnections also bear nonlinearities. The advanced strategy removes several difficulties found in analogous methodologies, namely eliminating control laws singularity issues, avoiding the use of Lyapunov functions with integrals, and suppressing the need for the introduction of virtual controllers¿ derivatives into the neural networks, which leads to less complex control designs, with decreased computational cost and with better response performance in comparison with similar existing methods / Doutorado / Automação / Doutor em Engenharia Elétrica

The replication of human intelligence, learning and reasoning by means of computer algorithms is termed Artificial Intelligence (Al) and the interaction of such algorithms with the physical world can be achieved using robotics. The work described in this thesis investigates the applications of concept learning (an approach which takes its inspiration from biological motivations and from survival instincts in particular) to robot control and path planning. The methodology of concept learning has been applied using learning decision trees (DTs) which induce domain knowledge from a finite set of training vectors which in turn describe systematically a physical entity and are used to train a robot to learn new concepts and to adapt its behaviour. To achieve behaviour learning, this work introduces the novel approach of hierarchical learning and knowledge decomposition to the frame of the reactive robot architecture. Following the analogy with survival instincts, the robot is first taught how to survive in very simple and homogeneous environments, namely a world without any disturbances or any kind of "hostility". Once this simple behaviour, named a primitive, has been established, the robot is trained to adapt new knowledge to cope with increasingly complex environments by adding further worlds to its existing knowledge. The repertoire of the robot behaviours in the form of symbolic knowledge is retained in a hierarchy of clustered decision trees (DTs) accommodating a number of primitives. To classify robot perceptions, control rules are synthesised using symbolic knowledge derived from searching the hierarchy of DTs. A second novel concept is introduced, namely that of multi-dimensional fuzzy associative memories (MDFAMs). These are clustered fuzzy decision trees (FDTs) which are trained locally and accommodate specific perceptual knowledge. Fuzzy logic is incorporated to deal with inherent noise in sensory data and to merge conflicting behaviours of the DTs. In this thesis, the feasibility of the developed techniques is illustrated in the robot applications, their benefits and drawbacks are discussed.