No description available.

A platoon of autonomous underwater vehicles provides a compelling platform for studying many challenging issues in multi-agent cooperative control. These challenges include developing cooperative algorithms suitable to practical multi-vehicle applications. They also include addressing intervehicle communication issues, such as sharing information via limited bandwidth channels and selecting network architecture to facilitate control design. This work addresses problems in each of these areas. Environmental tracking and formation control serves as the main application upon which this work focuses. In the tracking and formation control application, a team of vehicles obtains a spatial average of an environmental feature by collecting and sharing local measurements. To achieve this objective, vehicles track a desired environmental field contour with their average position while maintaining a desired spatial formation about the average. A decentralized consensus-based algorithm is developed for controlling the platoon. In a novel two-level consensus approach, each vehicle estimates a virtual leader trajectory using local and shared measurements at one level, then positions itself about the virtual leader at a second level. Due to very low bandwidth underwater communication, vehicles share information intermittently, and the platoon network is effectively disconnected at every instant of time. This issue is addressed by modeling the platoon as a periodic switched system whose frozen-time subsystems possess disconnected networks, but whose time-averaged system is connected. The stability and input-output properties of the switched system are related to those of the corresponding average system. Under sufficiently fast switching, asymptotic stability of the average system implies asymptotic stability of the switched system and the existence of an L2 gain. Estimates of the slowest stabilizing switching rate and the L2 gain are derived. Controller and estimator design are complicated by the lack of a separation principle for decentralized systems and by the effects of intervehicle coupling. The potential for choosing the communication topology in a manner that leads to design simplifications is investigated. In particular, a transformation is presented that converts the platoon state coefficient matrix to block diagonal form when the communication network has a circulant structure. / Ph. D.

Adaptive controllers allow robots to perform a wide variety of tasks, but the extensive computations required have generated an interest in developing decentralized adaptive controllers. Horner has designed an adaptive controller for a four-degree-of-freedom mobile robot and tested it through simulations. The study described in this thesis uses the techniques described by Horner to design and test a decentralized model reference adaptive controller (DMRAC) for a physical four-degree-of-freedom mobile robot. The study revealed several difficulties in implementing this design. Most notably, the robot available for the research did not allow for the measurement of joint velocity, so it was necessary to estimate the velocity as the derivative of the position measurement. The noise created by this estimation made completion of testing impossible. Future research should be performed on a robot that provides joint velocity measurement. Alternatively, a study could include state estimation as part of the controller, thus reducing and possibly eliminating the need for velocity measurement. / Master of Science

Control systems for robotic manipulators have been investigated for several years. The difficulty in designing a controller for a robotic manipulator is due to the highly nonlinear, time-varying dynamics. Closed-loop constant gain controllers are effective when the robot is expected to perform a limited range of operations. In the case of a mobile robot, the commanded tasks are not likely to be repetitive. Thus, another method of control is desired to overcome the effects of the nonlinear time-varying dynamics. Several adaptive control methods have been applied to robotic manipulators. The adaptive controllers are successful at trajectory tracking in the presence of the nonlinear time-varying dynamics. Some of these methods are computationally demanding, therefore, most of the current research focuses on efficient adaptive control methods. In particular, the area of decentralized adaptive control is gaining popularity. This method involves reducing a dynamic system into subsystems, each with an individual controller. This method is more efficient since the controllers can operate simultaneously. In this study, a decentralized model reference adaptive controller (MRAC) was designed for a four-degree-of-freedom mobile robot. The performance of the decentralized MRAC controller was compared to that of a constant gain state feedback controller. The decentralized MRAC control strategy proved to be an efficient method of control for a mobile robot that is superior to state feedback control when the robot is performing highly nonlinear time-varying tasks. Also, the computational load for each subsystem of the decentralized controller was less than the computational load of the state feedback controller. / Master of Science

Due to the ever-increasing penetration of distributed generation units connected to the power distribution system, electric power systems, worldwide, are undergoing a paradigm shift with regards to system monitoring, operation and control. We envision that with the emergence of `active' distribution systems consisting of `prosumers' and localized energy markets, decentralized control methods in power systems are gaining a growing attention among power researchers. Traditionally, two main types of control schemes have been implemented in power systems: (a) wide-area monitoring based centralized control, and (b) local measurement based primary (machine) level control. By contrast, decentralized control schemes based on local monitoring and control of strategically-determined subsystems (or `areas') of a large-scale power system are not used. The latter control schemes offer several advantages over the former, which include more flexibility, simplicity, economy and scalability for large-scale systems. In this dissertation, we summarize our research work on hierarchical and decentralized control techniques for the enhancement in a unified manner of voltage and rotor angle stability in large-scale power systems subject to large (e.g., short circuits) and small (e.g., small load changes) disturbances. We study system robustness by calculating local stability margins. We derive decentralized control laws that guaranty global asymptotic stability by applying Lyapunov's second method for interconnected systems. Furthermore, we argue that the current centralized control structure must only play a supervisory control role at a higher (tertiary) hierarchical level by processing the decisions taken by the regional control entities regarding the stability/instability of the system. This ensures system-wide situational awareness while minimizing the communication bandwidth requirements. We also develop a multi-agent based framework for this hierarchical control scheme. Finally, we compare different communication protocols using simulation models and propose an efficient communication network design for decentralized control schemes. This work, in principle, motivates the development of fast stability analysis which, in the future, may also account for the non-linear coupling that exist between machine rotor angles and bus voltages in power system models. As a future work, we propose the use of statistical techniques like random-effects regression and saddlepoint approximation method to reliably estimate the type-I and type-II probability errors in the proposed hierarchical, decentralized control decision process. / Ph. D.

Les travaux entrepris au cours de cette thèse ont permis de concevoir des stratégies de commande de systèmes multivariables (outils d’analyse et méthodes de synthèse) en vue de leur application au développement de lois de commande d’un turbopropulseur.D’un point de vue fonctionnel, un turbopropulseur est un système multivariable comprenant deux grandeurs de commande : le débit carburant à injecter dans la chambre de combustion et le pas de l’hélice, ainsi que deux grandeurs de sortie : la puissance délivrée par l’hélice et sa vitesse de rotation. Ces variables sont fortement couplées, ce qui signifie que des variations de l’une entraînent des écarts sur l’autre. L’objectif de ces travaux est de synthétiser des lois de commande facilement ajustables, permettant de respecter des spécifications classiques en Automatique (temps de réponse, dépassement, erreur statique) et de réduire les couplages entre les différentes grandeurs régulées. Dans ce contexte industriel, les approches décentralisées sans et avec découplage sont envisagées. La stratégie décentralisée pure met en œuvre un correcteur diagonal, ce qui revient à asservir un système par plusieurs boucles monovariables indépendantes. Bien que relativement facile à synthétiser et à implanter, la stratégie décentralisée ne permet pas d’atteindre les performances souhaitées en présence d’interactions importantes. Dans ce cas, il est possible de l’associer à des compensateurs permettant de diminuer les interactions.Une part importante de ces travaux de recherche concerne le développement méthodologique de ces stratégies. La définition d’une stratégie de commande est la première étape. Pour cela, la quantification du niveau d’interaction dans un système se révèle importante. Celle-ci peut être réalisée à l’aide de différentes méthodes et indicateurs qui s’appuient sur les réponses fréquentielles ou temporelles du système, ou encore sur les grammiens de commandabilité et d’observabilité. Une procédure systématique d’analyse des interactions a été proposée afin de déterminer la stratégie de commande la plus adaptée en fonction des interactions. Dans le cas où l’analyse des interactions conduit à adopter une stratégie décentralisée, les régulateurs peuvent être synthétisés à l’aide de méthodes monoboucles ou multiboucles. Les premières ne prennent pas en compte les interactions tandis que les secondes, plus élaborées mais également plus complexes à mettre en œuvre, permettent de les prendre spécifiquement en compte. A la suite de l’analyse de ces méthodes, une étude récapitulative présentant les méthodes préconisées en fonction du procédé et des objectifs, est finalement proposée. Dans le cas où l’analyse des interactions montre un niveau de couplage trop important, il est possible d’associer des compensateurs à la régulation décentralisée. Les compensateurs ont pour but de découpler les commandes vis-à-vis des sorties du procédé. Différentes méthodes et structures de découplage ont été étudiées et comparées. Une procédure de découplage, composée des méthodes considérées comme les plus efficaces a finalement été mise en place. / In this Ph.D. thesis, we explore the different steps of designing a decentralized control applied on a turboprop engine.From the control point of view, the turboprop engine is a TITO (Two-Input Two-Output) process. The fuel flow is used to control the shaft power while the blade pitch angle is used to control the propeller speed. The turboprop presents important couplings between manipulated variables and controlled variables. When the fuel flow changes, the propeller speed is impacted. Similarly, when controlling the blade pitch angle to change the propeller speed to another level, the shaft power is affected, particularly during the transient states. The main objective of this research thesis is to design control laws for the turboprop. Beside technical specifications like response time and overshoot, couplings between loops have to be reduced as much as possible and control laws have to be robust to model uncertainties. For this industrial environment a decentralized strategy (with or without compensators) has been chosen. The decentralized strategy consists in designing monoloop controllers in order to drive the multivariable system. The decentralized strategy presents important benefits, such as flexibility as well as design simplicity, but is not efficient in presence of heavy couplings. In that case, it is possible to use compensators that reduce existing process interactions before designing the monoloop controllers.An important part of this work focuses on the development of these different strategies.The first step is the choice of the control structure, which strongly depends on the level of interaction. Despite the availability of different metrics - based on frequential responses, temporal responses, or Gramian – it is not easy to know which one is the most appropriate. Based on the analysis of couplings with different metrics, a procedure is proposed in order to choose the structure and the controllers design method.If the coupling analysis leads to adopt a decentralized strategy, the controllers can be designed using monoloop or multiloop methods. The first ones are simple but do not take couplings into account, whereas multiloop methods take specifically couplings into account but are more complicated. These tuning methods have been studied and recommendations have finally been made to choose the most appropriated method depending on the process and the requirements.In cases where the couplings analysis reveals a high couplings level, compensators can be associated with the decentralized strategy. The objectives of the compensators are to reduce couplings in the system and to facilitate the design of monoloop controllers. Several decoupling structures have been studied and compared. A decoupling procedure has then been proposed.These methodological studies have been applied to the turboprop engine. Following the coupling analysis, a decentralized strategy with decoupling has been chosen. After following the decoupling procedure and trying different solutions, the inverted decoupler has been adopted. Considering the dynamics of the system and the total decoupling provided by the inverted decoupler, PI controllers have been used and a monoloop tuning method has been chosen. In order to guarantee the desired performances over the whole flight envelope, control laws have been interpolated, using a gain scheduling technique. The structured singular value approach has then been used to demonstrate the robustness of the control laws with model uncertainties. Control laws have finally been implemented in the control software and simulation results have illustrated their good performances.

Les réseaux sont présents dans plusieurs domaines scientifiques et d’ingénierie tels que la biologie, la physique, la sociologie ainsi que la robotique ou la théorie de la communication. L’étude de ces réseaux montre qu’ils sont souvent structurés en sous-groupes. Entre eux il n’y a pas ou il y a très peu d’interaction. Par conséquent, un accord local au sein de chaque groupe est naturellement atteint alors que le consensus associé à un tel réseau doit être imposé par une loi de commande spécifique. Nous proposons donc un contrôleur discret quasi-périodique pour échanger des informations entre les groupes. Un agent de chaque groupe est choisi le « leader » et, à certains moments, ces leaders communiquent entre eux à travers un nouveau réseau. Ceci permet d’obtenir le consensus dans tout le réseau mais engendre des réinitialisation/sauts dans l’état de leaders. La première contribution de la thèse est la caractérisation de la valeur de consensus dans le cadre des systèmes linéaires impulsifs. Il est remarquable que la valeur de consensus dépende seulement des conditions initiales et des topologies des réseaux impliqués. Elle n’est donc pas sensible aux instants de réinitialisation des états de leaders. Afin d’étudier la stabilité de la valeur de consensus obtenue, nous proposons une méthode fondée sur la vérification d’une condition LMI. Cela peut être adaptée pour la conception du réseau d’interaction entre les leaders permettant d’atteindre une valeur de consensus a priori choisie. Il est aussi possible d’utiliser la condition LMI afin de garantir une vitesse de convergence désirée vers le consensus. Pour ces derniers objectifs, la topologie du réseau continue à être considérée comme fixe et connue pour chaque groupe. L’ensemble des valeurs de consensus qui peuvent être atteintes est contenu dans l’intervalle défini par le minimum et le maximum des accords locaux initiaux. Ensuite nous présentons l’étude d’un problème pratique. Des robots mobiles non-holonome, séparés dans des groupes, doivent atteindre une formation donnée. L’algorithme de consensus à pour mission de définir les trajectoires de référence pour ces robots en prenant en compte juste les informations locale. Le robot poursuit la trajectoire de référence en utilisant une commande classique pour cela. / Networks appear in several areas of science and engineering such as biology, physics, sociology as well as robotics and communication theory. Studying these networks it is possible to see cluster-like structures, which are disconnected or very weakly connected one to another. The presence of these clusters hampers consensus throughout the overall network. Instead, local agreement is reached within each cluster. To enforce consensus we have to design an appropriate decentralized controller that imposes interactions between clusters. While the interactions inside each cluster are continuous, we propose a quasi-periodic discrete controller to exchange information between clusters. A single agent from each cluster is chosen to be the leader, and at certain moments, the leaders communicate with each other through a new network. This allows consensus in the entire network but generates resets/jumps on the leaders’ state. The first contribution of this manuscript is related to the characterization of the consensus value in the linear impulsive dynamics framework. It is noteworthy that the consensus value depends only on the initial conditions and the topologies of the involved networks. Therefore, the consensus value does not depend on the reset sequence used for the leaders’ states. To study the stability of the consensus value a LMI based condition is proposed. The main advantage of this approach is its flexibility. Indeed with some modifications to the LMI condition it is possible to analyse the convergence speed of the network or to design the leaders’ network. The purpose of leaders’ network design is to reach an a priori specified consensus value with a specified convergence speed. Whatever is the objective, throughout the manuscript we consider that the network topology is fixed and known for each cluster. The set of consensus values that can be reached is restricted to the interval defined by the minimum and maximum initial local agreements. A last contribution is related to the application of the proposed methodology to a practical situation. We consider a fleet of non-holonomic mobile robots separated in clusters. The communication inside each cluster are secured and cheap while between clusters it is expensive and not securely to communicate. Nevertheless the robots have to reach a given formation. In this case our consensus algorithm is in charge of providing reference trajectories to each robot by using only the available local information. The robot follows the reference by using a classical trajectory tracking control.

Les travaux entrepris au cours de cette thèse ont permis de concevoir des stratégies de commande de systèmes multivariables (outils d’analyse et méthodes de synthèse) en vue de leur application au développement de lois de commande d’un turbopropulseur.D’un point de vue fonctionnel, un turbopropulseur est un système multivariable comprenant deux grandeurs de commande : le débit carburant à injecter dans la chambre de combustion et le pas de l’hélice, ainsi que deux grandeurs de sortie : la puissance délivrée par l’hélice et sa vitesse de rotation. Ces variables sont fortement couplées, ce qui signifie que des variations de l’une entraînent des écarts sur l’autre. L’objectif de ces travaux est de synthétiser des lois de commande facilement ajustables, permettant de respecter des spécifications classiques en Automatique (temps de réponse, dépassement, erreur statique) et de réduire les couplages entre les différentes grandeurs régulées. Dans ce contexte industriel, les approches décentralisées sans et avec découplage sont envisagées. La stratégie décentralisée pure met en œuvre un correcteur diagonal, ce qui revient à asservir un système par plusieurs boucles monovariables indépendantes. Bien que relativement facile à synthétiser et à implanter, la stratégie décentralisée ne permet pas d’atteindre les performances souhaitées en présence d’interactions importantes. Dans ce cas, il est possible de l’associer à des compensateurs permettant de diminuer les interactions.Une part importante de ces travaux de recherche concerne le développement méthodologique de ces stratégies. La définition d’une stratégie de commande est la première étape. Pour cela, la quantification du niveau d’interaction dans un système se révèle importante. Celle-ci peut être réalisée à l’aide de différentes méthodes et indicateurs qui s’appuient sur les réponses fréquentielles ou temporelles du système, ou encore sur les grammiens de commandabilité et d’observabilité. Une procédure systématique d’analyse des interactions a été proposée afin de déterminer la stratégie de commande la plus adaptée en fonction des interactions. Dans le cas où l’analyse des interactions conduit à adopter une stratégie décentralisée, les régulateurs peuvent être synthétisés à l’aide de méthodes monoboucles ou multiboucles. Les premières ne prennent pas en compte les interactions tandis que les secondes, plus élaborées mais également plus complexes à mettre en œuvre, permettent de les prendre spécifiquement en compte. A la suite de l’analyse de ces méthodes, une étude récapitulative présentant les méthodes préconisées en fonction du procédé et des objectifs, est finalement proposée. Dans le cas où l’analyse des interactions montre un niveau de couplage trop important, il est possible d’associer des compensateurs à la régulation décentralisée. Les compensateurs ont pour but de découpler les commandes vis-à-vis des sorties du procédé. Différentes méthodes et structures de découplage ont été étudiées et comparées. Une procédure de découplage, composée des méthodes considérées comme les plus efficaces a finalement été mise en place. / In this Ph.D. thesis, we explore the different steps of designing a decentralized control applied on a turboprop engine.From the control point of view, the turboprop engine is a TITO (Two-Input Two-Output) process. The fuel flow is used to control the shaft power while the blade pitch angle is used to control the propeller speed. The turboprop presents important couplings between manipulated variables and controlled variables. When the fuel flow changes, the propeller speed is impacted. Similarly, when controlling the blade pitch angle to change the propeller speed to another level, the shaft power is affected, particularly during the transient states. The main objective of this research thesis is to design control laws for the turboprop. Beside technical specifications like response time and overshoot, couplings between loops have to be reduced as much as possible and control laws have to be robust to model uncertainties. For this industrial environment a decentralized strategy (with or without compensators) has been chosen. The decentralized strategy consists in designing monoloop controllers in order to drive the multivariable system. The decentralized strategy presents important benefits, such as flexibility as well as design simplicity, but is not efficient in presence of heavy couplings. In that case, it is possible to use compensators that reduce existing process interactions before designing the monoloop controllers.An important part of this work focuses on the development of these different strategies.The first step is the choice of the control structure, which strongly depends on the level of interaction. Despite the availability of different metrics - based on frequential responses, temporal responses, or Gramian – it is not easy to know which one is the most appropriate. Based on the analysis of couplings with different metrics, a procedure is proposed in order to choose the structure and the controllers design method.If the coupling analysis leads to adopt a decentralized strategy, the controllers can be designed using monoloop or multiloop methods. The first ones are simple but do not take couplings into account, whereas multiloop methods take specifically couplings into account but are more complicated. These tuning methods have been studied and recommendations have finally been made to choose the most appropriated method depending on the process and the requirements.In cases where the couplings analysis reveals a high couplings level, compensators can be associated with the decentralized strategy. The objectives of the compensators are to reduce couplings in the system and to facilitate the design of monoloop controllers. Several decoupling structures have been studied and compared. A decoupling procedure has then been proposed.These methodological studies have been applied to the turboprop engine. Following the coupling analysis, a decentralized strategy with decoupling has been chosen. After following the decoupling procedure and trying different solutions, the inverted decoupler has been adopted. Considering the dynamics of the system and the total decoupling provided by the inverted decoupler, PI controllers have been used and a monoloop tuning method has been chosen. In order to guarantee the desired performances over the whole flight envelope, control laws have been interpolated, using a gain scheduling technique. The structured singular value approach has then been used to demonstrate the robustness of the control laws with model uncertainties. Control laws have finally been implemented in the control software and simulation results have illustrated their good performances.

The behaviour of a large nuclear reactor can be described with sufficient accuracy using a nodal model, like the spatial model of a 540 MWe large Pressurized Heavy Water Reactor (PHWR). This model divides the reactor into divisions or nodes to create a spatial model in order to control the xenon induced oscillations that occur in PHWRs. However, being such a large scale system, a 72nd-order model, it makes controller design challenging. Therefore, a reduced order model is much more manageable. A convenient method of model reduction while maintaining the important dynamics characteristics of the process can be done by decoupling. Also, due to the nature of the system, decentralized controllers could serve as a better option because it allows each controller to be localized. This way, any control input to a zone only affects the desired zone and the zones most coupled with, thus not causing a respective change in neutron flux in the other zones. In this thesis, three decentralized controllers were designed using the spatial model of a 540 MWe large PHWR. A decoupling algorithm was designed to divide the system into three partitions containing 20, 27, and 25 states each. Reduced order sub-systems were thus created to produce optimal decentralized controllers. An optimal centralized controller was created to compare both approaches. The decentralized versus centralized controllers’ system responses were analyzed after a reactivity disturbance. A fail-safe study was done to highlight one of the advantages of decentralized controllers. / UOIT

This thesis presents decentralized model reference adaptive control techniques for systems with full-state feedback and systems with output feedback. The controllers are strictly decentralized, that is, each local controller uses feedback from only local subsystems and no information is shared between local controllers. The full-state feedback decentralized controller is effective for multi-input systems, where the dynamics matrix and control-input matrix are unknown. The decentralized controller achieves asymptotic stabilization and command following in the presence of sinusoidal disturbances with known spectrum. We present a construction technique of the reference-model dynamics such that the decentralized controller is effective for systems with arbitrarily large subsystem interconnections. The output-feedback decentralized controller is effective for single-input single-output subsystems that are minimum phase and relative degree one. The decentralized controller achieves asymptotic stabilization and disturbance rejection in the presence of an unknown disturbance, which is generated by an unknown Lyapunov-stable linear system.