Designing local controllers for networked systems is challenging, because in these systems each local controller can often access only part of the overall information on system parameters and sensor measurements. Traditional control design cannot be easily applied due to the unconventional information patterns, communication network imperfections, and design procedure complexities. How to control large-scale systems is of immediate societal importance as they appear in many emerging applications, such as intelligent transportation systems, smart grids, and energy-efficient buildings. In this thesis, we make three contributions to the problem of designing networked controller under information asymmetries and limitations. In the first contribution, we investigate how to design local controllers to optimize a cost function using only partial knowledge of the model governing the system. Specifically, we derive some fundamental limitations in the closed-loop performance when the design of each controller only relies on local plant model information. Results are characterized in the structure of the networked system as well as in the available model information. Both deterministic and stochastic formulations are considered for the closed-loop performance and the available information. In the second contribution of the thesis, we study decision making in transportation systems using heterogeneous routing and congestion games. It is shown that a desirable global behavior can emerge from simple local strategies used by the drivers to choose departure times and routes. Finally, the third contribution is a novel stochastic sensor scheduling policy for ad-hoc networked systems, where a varying number of control loops are active at any given time. It is shown that the policy provides stochastic guarantees for the network resources dynamically allocated to each loop. / <p>QC 20140221</p>

Ο σκοπός αυτής της διπλωματικής εργασίας είναι να συζητηθούν και να παρουσιαστούν οι υπάρχουσες τεχνικές που εφαρμόζουν την στρατηγική δειγματοληψίας βασισμένη σε γεγονότα για την εκτίμηση της κατάστασης των γραμμικών συστημάτων και να αξιολογηθεί πώς αυτές οι μέθοδοι επηρεάζουν τη συνολική απόδοση του συστήματος. Όπως ήδη αναφέρθηκε, περιορισμοί στον ρυθμό δειγματοληψίας ανακύπτουν στα NCS λόγω του περιορισμένου διαθέσιμου εύρους ζώνης. Αυτός ο περιορισμός στον αριθμό δειγμάτων επηρεάζει σαφώς το μέσο τετραγωνικό σφάλμα εκτίμησης, και η ερώτηση είναι πώς πρέπει να επιλέξουμε τις χρονικές στιγμές δειγματοληψίας προκειμένου να το ελαχιστοποιήσουμε. / The aim of this diploma thesis is to discuss and present existing techniques that apply event-triggered sampling to linear system state estimation, and to evaluate how these methods affect the overall performance of the system. As already mentioned, sampling rate constraints arise in NCS due to the limited bandwidth available. This restriction on the number of samples clearly affects the mean square estimation distortion error, and the question is how we should choose the sampling instants in order to minimize it.

<p>Distributed optimization methods allow us to decompose an optimization problem</p><p>into smaller, more manageable subproblems that are solved in parallel. For this</p><p>reason, they are widely used to solve large-scale problems arising in areas as diverse</p><p>as wireless communications, optimal control, machine learning, artificial intelligence,</p><p>computational biology, finance and statistics, to name a few. Moreover, distributed</p><p>algorithms avoid the cost and fragility associated with centralized coordination, and</p><p>provide better privacy for the autonomous decision makers. These are desirable</p><p>properties, especially in applications involving networked robotics, communication</p><p>or sensor networks, and power distribution systems.</p><p>In this thesis we propose the Accelerated Distributed Augmented Lagrangians</p><p>(ADAL) algorithm, a novel decomposition method for convex optimization prob-</p><p>lems with certain separability structure. The method is based on the augmented</p><p>Lagrangian framework and addresses problems that involve multiple agents optimiz-</p><p>ing a separable convex objective function subject to convex local constraints and</p><p>linear coupling constraints. We establish the convergence of ADAL and also show</p><p>that it has a worst-case O(1/k) convergence rate, where k denotes the number of</p><p>iterations.</p><p>Moreover, we show that ADAL converges to a local minimum of the problem</p><p>for cases with non-convex objective functions. This is the first published work that</p><p>formally establishes the convergence of a distributed augmented Lagrangian method</p><p>ivfor non-convex optimization problems. An alternative way to select the stepsizes</p><p>used in the algorithm is also discussed. These two contributions are independent</p><p>from each other, meaning that convergence of the non-convex ADAL method can</p><p>still be shown using the stepsizes from the convex case, and, similarly, convergence</p><p>of the convex ADAL method can be shown using the stepsizes proposed in the non-</p><p>convex proof.</p><p>Furthermore, we consider cases where the distributed algorithm needs to operate</p><p>in the presence of uncertainty and noise and show that the generated sequences of</p><p>primal and dual variables converge to their respective optimal sets almost surely. In</p><p>particular, we are concerned with scenarios where: i) the local computation steps</p><p>are inexact or are performed in the presence of uncertainty, and ii) the message</p><p>exchanges between agents are corrupted by noise. In this case, the proposed scheme</p><p>can be classified as a distributed stochastic approximation method. Compared to</p><p>existing literature in this area, our work is the first that utilizes the augmented</p><p>Lagrangian framework. Moreover, the method allows us to solve a richer class of</p><p>problems as compared to existing methods on distributed stochastic approximation</p><p>that consider only consensus constraints.</p><p>Extensive numerical experiments have been carried out in an effort to validate</p><p>the novelty and effectiveness of the proposed method in all the areas of the afore-</p><p>mentioned theoretical contributions. We examine problems in convex, non-convex,</p><p>and stochastic settings where uncertainties and noise affect the execution of the al-</p><p>gorithm. For the convex cases, we present applications of ADAL to certain popular</p><p>network optimization problems, as well as to a two-stage stochastic optimization</p><p>problem. The simulation results suggest that the proposed method outperforms</p><p>the state-of-the-art distributed augmented Lagrangian methods that are known in</p><p>the literature. For the non-convex cases, we perform simulations on certain simple</p><p>non-convex problems to establish that ADAL indeed converges to non-trivial local</p><p>vsolutions of the problems; in comparison, the straightforward implementation of the</p><p>other distributed augmented Lagrangian methods on the same problems does not</p><p>lead to convergence. For the stochastic setting, we present simulation results of</p><p>ADAL applied on network optimization problems and examine the effect that noise</p><p>and uncertainties have in the convergence behavior of the method.</p><p>As an extended and more involved application, we also consider the problem</p><p>of relay cooperative beamforming in wireless communications systems. Specifically,</p><p>we study the scenario of a multi-cluster network, in which each cluster contains</p><p>multiple single-antenna source destination pairs that communicate simultaneously</p><p>over the same channel. The communications are supported by cooperating amplify-</p><p>and-forward relays, which perform beamforming. Since the emerging problem is non-</p><p>convex, we propose an approximate convex reformulation. Based on ADAL, we also</p><p>discuss two different ways to obtain a distributed solution that allows for autonomous</p><p>computation of the optimal beamforming decisions by each cluster, while taking into</p><p>account intra- and inter-cluster interference effects.</p><p>Our goal in this thesis is to advance the state-of-the-art in distributed optimization by proposing methods that combine fast convergence, wide applicability, ease</p><p>of implementation, low computational complexity, and are robust with respect to</p><p>delays, uncertainty in the problem parameters, noise corruption in the message ex-</p><p>changes, and inexact computations.</p> / Dissertation

Este trabalho apresenta uma proposta de solução para compensar o atraso variável, característico de redes IP. O preditor de Smith é tradicionalmente usado para compensar atrasos, mas no caso deste tipo de rede um fator importante que pode degradar o controle do sistema são as grandes variações nos atrasos sofridos pelas mensagens. Para tratar este problema, este trabalho baseia-se na criação de uma estrutura de dados de armazenamento para a saída do preditor de Smith, possibilitando a comparação do valor de saída da planta com o valor adequado guardado na estrutura. Para determinação do valor correspondente, este está associado a um índice calculado com o tempo de roundtrip obtido de informações dos pacotes de dados proveniente da planta. Para se chegar a tal solução, foi analisada nas primeiras seções a influência do atraso em malhas de controle e posteriormente um estudo do estado da arte para tipos de controle sobre redes IP. Ao final é apresentada uma análise experimental, demonstrando resultados promissores da aplicação desta proposta. / This work presents a solution to compensate the varying delay, characteristic of IP networks. The Smith predictor is traditionally used to compensate delays but in this type of networks an other important factor that can destabilize the system is the large jitter of the delay. In order to solve this problem, the solution presented here is based on a data structure to save the output from the model of the plant used by the Smith predictor, that make possible to compare the output from the plant with the corresponding value in the structure. In order to take the correspondent value, it is associated to an index which is calculated with the roundtrip obtained from the informations in the packet arrived from the plant. In order to formulate this solution, it was analised in the first sections the influence of the delay in control systems and then a study of the state of art of the control systems over IP networks. At the end, an experimental analise to demonstrate the good result of the application of this solution is presented.

Este trabalho aborda o problema de controle baseado em eventos para sistemas em tempo discreto, considerando que o sistema possui os dispositivos atuadores e sensores em nós diferentes e separados por uma rede de comunicação. A estratégia baseada em eventos consiste em reduzir a utilização da rede ao transmitir as informações do sensor para o atuador apenas quando um evento é gerado pela violação de um determinado limiar pela função de disparo. Primeiramente, são formuladas condições para a estabilidade de um sistema linear com realimentação estática de estados sob a estratégia proposta, com base na teoria de Lyapunov. Como as condições são postas na forma de desigualdades matriciais lineares (LMIs, do inglês linear matrix inequalities), problemas de otimização convexos podem ser utilizados na determinação dos parâmetros da função de disparo, bem como na resolução do problema de co-design, ou seja, do projeto simultâneo do controlador e da função de disparo, os quais são providos na sequência. A partir deste resultado básico, a metodologia é estendida para o caso em que ocorre a saturação do atuador. A seguir, é apresentada a extensão da metodologia para o caso em que o estado da planta não está disponível para o sensor, sendo então utlizado um observador de estados, considerando-se tanto o caso em que o modelo da planta utilizado no observador corresponde exatamente à dinâmica real da planta quanto o caso em que este modelo apresenta incertezas. Exemplos numéricos são apresentados para ilustrar todas as classes de sistemas consideradas, com os quais constata-se que a estratégia proposta é eficiente na redução da utilização dos recursos da rede de comunicação. / This work approaches the problem of event-triggered control for discrete time systems, considering that the system has the actuator and sensor devices in different nodes, separated by a communication network. The event-triggered strategy consists in reducing the utilization of the network by only transmitting the information from the sensor to the actuator when an event is generated by the violation of a determined threshold by the trigger function. Firstly, conditions for the stability of a linear system with a static state feedback under the proposed strategy are formulated based on the Lyapunov theory. Since the conditions are given in the form of linear matrix inequalities (LMIs), convex optimization problems can be used for the determination of the trigger function parameters, as well as the co-design of the feedback gain and the trigger function, which are given next. From this basic result, the methodology is extended to the case where occurs the saturation of the actuator. Following, the extension of the methodlogy to the case in which the plant states are not available for measure is presented, and a state-observer is used, considering both the case that the plant model corresponds exactly to the real plant dynamics and the case where this model has uncertainties. Numeric examples are shown to illustrate all the system classes considered, with which it is found that the proposed strategy is efficient in the reduction of the network resources utilization.

A importância das redes de comunicação industriais em modernos sistemas de automação e controle industriais tem aumentado significativamente nos últimos anos, devido aos avanços nas áreas de processadores e softwares embarcados, que permitem o desenvolvimento de dispositivos com elevada capacidade de processamento a custos reduzidos. Estas características também são muito importantes em sistemas automotivos, visto que existe uma tendência para a substituição de sistemas mecânicos e hidráulicos em veículos e o espaço disponível para implementação é bastante reduzido. Esta substituição passa pela elaboração de complexos algoritmos de controle, os quais, quando operam sobre uma rede de comunicação, precisam considerar explicitamente os efeitos do canal de comunicação compartilhado na dinâmica do sistema em malha fechada. Este trabalho apresenta uma análise do impacto da comunicação em rede sobre sistemas de controle. Mais especificamente, analisa-se o comportamento do protocolo Flexray, recentemente desenvolvido por um consórcio de importantes empresas e que incorpora interessantes conceitos para escalonamento de mensagens síncronas e assíncronas. No trabalho foram realizados experimentos com três diferentes tipos de controladores aplicados ao estudo de caso de uma suspensão ativa, onde o sistema tem sua malha fechada sobre a rede FlexRay. / The importance of communication networks on modern automation systems has increased significantly over the last years, mostly due to advances in embedded microprocessor and software technologies, which enable the development of devices with high processing power at reduced costs. These characteristics are very important for vehicle systems, since there is nowadays a trend to replace mechanical and hydraulic systems, and the space available for implementation is limited. This replacement requires very complex control algorithms, which, when operating on a communication network, have to consider explicitly the effects introduced by the shared communication channel on the closed loop system dynamics. This work presents an analysis of the network communication impact over control systems. More specifically, it is of interest to analyse the behavior of the FlexRay protocol, which has been recently developed by a Consortium of important companies and incorporates interesting concepts of synchronous and asynchronous message scheduling. In this work, some experiments were performed with three controllers, which were developed using different methodologies, applied to the case study of an active suspension system, where the loop is closed over the FlexRay protocol.

Este trabalho apresenta uma proposta de solução para compensar o atraso variável, característico de redes IP. O preditor de Smith é tradicionalmente usado para compensar atrasos, mas no caso deste tipo de rede um fator importante que pode degradar o controle do sistema são as grandes variações nos atrasos sofridos pelas mensagens. Para tratar este problema, este trabalho baseia-se na criação de uma estrutura de dados de armazenamento para a saída do preditor de Smith, possibilitando a comparação do valor de saída da planta com o valor adequado guardado na estrutura. Para determinação do valor correspondente, este está associado a um índice calculado com o tempo de roundtrip obtido de informações dos pacotes de dados proveniente da planta. Para se chegar a tal solução, foi analisada nas primeiras seções a influência do atraso em malhas de controle e posteriormente um estudo do estado da arte para tipos de controle sobre redes IP. Ao final é apresentada uma análise experimental, demonstrando resultados promissores da aplicação desta proposta. / This work presents a solution to compensate the varying delay, characteristic of IP networks. The Smith predictor is traditionally used to compensate delays but in this type of networks an other important factor that can destabilize the system is the large jitter of the delay. In order to solve this problem, the solution presented here is based on a data structure to save the output from the model of the plant used by the Smith predictor, that make possible to compare the output from the plant with the corresponding value in the structure. In order to take the correspondent value, it is associated to an index which is calculated with the roundtrip obtained from the informations in the packet arrived from the plant. In order to formulate this solution, it was analised in the first sections the influence of the delay in control systems and then a study of the state of art of the control systems over IP networks. At the end, an experimental analise to demonstrate the good result of the application of this solution is presented.

L'énergie est une ressource clé dans les systèmes commandés en réseau, en particulier dans les applications concernant les réseaux sans fil. Cette thèse étudie comment économiser l'énergie dans les capteurs sans fil avec une co-Conception contrôle et communication. Cette thèse examine les techniques et les approches existantes qui sont utilisées pour économiser l'énergie d'un point de vue de la communication et du contrôle. Cet étude est organisée selon une architecture de communication par couches couvrant de bas en haut les couches Physique, Liaison, Réseau, et Application. Puis, à partir de la conclusion que la puce radio est un important consommateur d'énergie, une loi conjointe de gestion des modes radio et de contrôle en boucle fermée est établie. La gestion des modes radio exploite les capacités de la puce radio à communter dans des modes de basses consommation pour économiser l'énergie, et d'adapter la puissance de transmission aux conditions du canal. Il en résulte un système de contrôle basé sur des événements où le système fonctionne en boucle ouverte à certains moments. Un compromis naturel apparaît entre l'économie d'énergie et les performances de contrôle. La loi conjointe est établie avec une formulation de contrôle optimal utilisant la Programmation Dynamique. Cette thèse résout le problème optimal dans les deux cas d'horizon infini et fini. La stabilité du système en boucle fermée est étudiée avec la formulation Input-To-State Stability (ISS). La principale conclusion de cette thèse, également illustrée dans la simulation, est que la conception à travers différentes couches dans les systèmes commandés en réseau est essentielle pour économiser l'énergie dans les noeuds sans fil. / Energy is a key resource in Networked Control Systems, in particular in applications concerning wireless networks. This thesis investigates how to save energy in wireless sensor nodes with control and communication co-Design. This thesis reviews existing techniques and approaches that are used to save energy from a communication and a control point of view. This review is organized according to the layered communication architecture covering from bottom to top the Physical, Data Link, Network, and Application layers. Then, from the conclusion that the radio chip is an important energy consumer, a joint radio-Mode management and feedback law policy is derived. The radio-Mode management exploits the capabilities of the radio chip to switch to low consuming radio-Modes to save energy, and to adapt the transmission power to the channel conditions. This results in an event-Based control scheme where the system runs open loop at certain time. A natural trade-Off appears between energy savings and control performance. The joint policy is derived in the framework of Optimal Control with the use of Dynamic Programming. This thesis solves the optimal problem in both infinite and finite horizon cases. Stability of the closed loop system is investigated with Input-To-State Stability framework. The main conclusion of this thesis, also shown in simulation, is that cross-Layer design in Networked Control System is essential to save energy in the wireless nodes.

Submitted by Marcos Rogerio Caldieri null (mrcaldieri@gmail.com) on 2016-12-14T23:11:29Z No. of bitstreams: 1 Marcos-Mestrado.pdf: 2168690 bytes, checksum: 11f237fc5587a57486d006c939b576af (MD5) / Approved for entry into archive by Felipe Augusto Arakaki (arakaki@reitoria.unesp.br) on 2016-12-19T17:40:14Z (GMT) No. of bitstreams: 1 caldieri_mr_me_bauru.pdf: 2168690 bytes, checksum: 11f237fc5587a57486d006c939b576af (MD5) / Made available in DSpace on 2016-12-19T17:40:14Z (GMT). No. of bitstreams: 1 caldieri_mr_me_bauru.pdf: 2168690 bytes, checksum: 11f237fc5587a57486d006c939b576af (MD5) Previous issue date: 2016-10-21 / A recente introdução de transmissores sem fio na indústria provocou um novo interesse em técnicas de medição e controle, porém a maioria das aplicações está restrita a medições de variáveis de processo em malha aberta ou aplicações de monitoramento. O motivo é a falta de confiabilidade devido aos problemas inerentes ao meio de transmissão, que pode ser a perda de pacotes de informação, atrasos de comunicação variantes no tempo, atualização muito lenta e não periódica da medição e vários tipos de interferências. A maior parte dos controladores industriais em controle de processos assumem que o ciclo de controle é executado de forma periódica e que uma nova medição está disponível para ser usada em intervalos de tempo conhecidos. No entanto esta situação não pode ser garantida quando sensores ou transmissores sem fio são usados em aplicações de controle em malha fechada, denominadas de Sistemas de Controle via Redes sem fio (WNCS – Wireless Networked Control Systems). Nesses tipos de aplicações, os transmissores sem fio devem transmitir novas medições de forma não periódica e somente se a medição da variável do processo tiver alterado significativamente. Para tornar esta tecnologia de WNCS mais confiável, muitas técnicas de controle têm sido pesquisadas, entre elas o PIDPlus que representa uma modificação do algoritmo PID para controle via rede sem fio. Este trabalho apresenta a implementação do protocolo Modbus para aplicações de WNCS. O protocolo Modbus TCP foi embarcado em hardware dedicado viabilizando a transmissão de dados via Ethernet TCP/IP e Wi-Fi. Uma comparação e avaliação de controladores PID para aplicação em WNCS sob condições de amostragem e atrasos de comunicações variáveis e de perdas de transmissão de mensagens foi realizada. Os resultados são analisados do ponto de vista de desempenho de controle e robustez. Resultados experimentais numa planta piloto comprovam a eficiência da implementação de uma malha de controle sem fio usando uma rede Wi-Fi com o protocolo Modbus embarcado e um controlador PIDPlus. / The recent introduction of wireless transmitters in the industry has driven a new interest in measuring and control techniques, but most applications are restricted to measurements of process variables in open loop or monitoring applications. The reason is the lack of reliability due to problems inherent to the transmission medium, which may be the packet loss, time varying delay, slow and aperiodic measurement updates and interference. Most industrial process controllers assume that the control cycle is performed periodically and that a new measurement is available to be used at known time intervals. However it cannot be guaranteed when wireless sensors or transmitters are used in closed loop control applications, called Wireless Networked Control Systems (WNCS). In these type of applications, wireless transmitters shall transmit new measurements not periodically and only if the process variable measurement has changed significantly. In order to enable and make this WNCS technology reliable, many control techniques have been researched including the PIDPlus that is a modified PID algorithm for wireless control. This paper presents the implementation of the Modbus protocol for WNCS applications. The Modbus TCP was embedded in dedicated hardware enabling the transmission of data via Ethernet TCP/IP and Wi-Fi. A comparison and evaluation of PID controllers for WNCS were done considering situations of variable sampling and communication delays and packet losses. The results are analyzed from the point of view of control performance and robustness. Experimental results in a pilot plant prove the efficiency of the implementation of a wireless control loop using a Wi-Fi network with embedded Modbus protocol and PIDPlus controller.

A importância das redes de comunicação industriais em modernos sistemas de automação e controle industriais tem aumentado significativamente nos últimos anos, devido aos avanços nas áreas de processadores e softwares embarcados, que permitem o desenvolvimento de dispositivos com elevada capacidade de processamento a custos reduzidos. Estas características também são muito importantes em sistemas automotivos, visto que existe uma tendência para a substituição de sistemas mecânicos e hidráulicos em veículos e o espaço disponível para implementação é bastante reduzido. Esta substituição passa pela elaboração de complexos algoritmos de controle, os quais, quando operam sobre uma rede de comunicação, precisam considerar explicitamente os efeitos do canal de comunicação compartilhado na dinâmica do sistema em malha fechada. Este trabalho apresenta uma análise do impacto da comunicação em rede sobre sistemas de controle. Mais especificamente, analisa-se o comportamento do protocolo Flexray, recentemente desenvolvido por um consórcio de importantes empresas e que incorpora interessantes conceitos para escalonamento de mensagens síncronas e assíncronas. No trabalho foram realizados experimentos com três diferentes tipos de controladores aplicados ao estudo de caso de uma suspensão ativa, onde o sistema tem sua malha fechada sobre a rede FlexRay. / The importance of communication networks on modern automation systems has increased significantly over the last years, mostly due to advances in embedded microprocessor and software technologies, which enable the development of devices with high processing power at reduced costs. These characteristics are very important for vehicle systems, since there is nowadays a trend to replace mechanical and hydraulic systems, and the space available for implementation is limited. This replacement requires very complex control algorithms, which, when operating on a communication network, have to consider explicitly the effects introduced by the shared communication channel on the closed loop system dynamics. This work presents an analysis of the network communication impact over control systems. More specifically, it is of interest to analyse the behavior of the FlexRay protocol, which has been recently developed by a Consortium of important companies and incorporates interesting concepts of synchronous and asynchronous message scheduling. In this work, some experiments were performed with three controllers, which were developed using different methodologies, applied to the case study of an active suspension system, where the loop is closed over the FlexRay protocol.