Asymptotic stability is a key notion of system stability for controlled dynamical systems as it guarantees that the system trajectories are bounded in a neighborhood of a given isolated equilibrium point and converge to this equilibrium over the infinite horizon. In some applications, however, asymptotic stability is not the appropriate notion of stability. For example, for systems with a continuum of equilibria, every neighborhood of an equilibrium contains another equilibrium and a nonisolated equilibrium cannot be asymptotically stable. Alternatively, in stabilization of spacecraft dynamics via gimballed gyroscopes, it is desirable to find state- and output-feedback control laws that guarantee partial-state stability of the closed-loop system, that is, stability with respect to part of the system state. Furthermore, we may additionally require finite-time stability of the closed-loop system, that is, convergence of the system's trajectories to a Lyapunov stable equilibrium in finite time. The Hamilton-Jacobi-Bellman optimal control framework provides necessary and sufficient conditions for the existence of state-feedback controllers that minimize a given performance measure and guarantee asymptotic stability of the closed-loop system. In this research, we provide extensions of the Hamilton-Jacobi-Bellman optimal control theory to develop state-feedback control laws that minimize nonlinear-nonquadratic performance criteria and guarantee semistability, partial-state stability, finite-time stability, and finite-time partial state stability of the closed-loop system.

This thesis considers the application of singular value decomposition (SVD) and semi-nonnegative matrix factorization (SNMF) within feedback control systems, called the SVD System and SNMF System, to control numerous subsystems with a reduced number of control inputs. The subsystems are coupled using a row-column structure to allow mn subsystems to be controlled using m+n inputs. Past techniques for controlling systems in this row-column structure have focused on scheduling procedures that offer limited performance. The SVD and SNMF Systems permit simultaneous control of every subsystem, which increases the convergence rate by an order of magnitude compared with previous methods. In addition to closed loop control, open loop procedures using the SVD and SNMF are compared with previous scheduling procedures, demonstrating significant performance improvements. This thesis presents theoretical results for the controllability of systems using the row-column structure and for the stability and performance of the SVD and SNMF Systems. Practical challenges to the implementation of the SVD and SNMF Systems are also examined. Numerous simulation examples are provided, in particular, a dynamic simulation of a pin array device, called Digital Clay, and two physical demonstrations are used to assess the feasibility of the SVD and SNMF Systems for specific applications.

Resilient networks have the ability to provide the desired level of service, despite challenges such as malicious attacks and misconfigurations. The primary goal of this dissertation is to be able to provide uninterrupted network services in the face of an attack or any failures. This dissertation attempts to apply control system theory techniques with a focus on system identification and closed-loop feedback control. It explores the benefits of system identification technique in designing and validating the model for the complex and dynamic networks. Further, this dissertation focuses on designing robust feedback control mechanisms that are both scalable and effective in real-time. It focuses on employing dynamic and predictive control approaches to reduce the impact of an attack on network services. The closed-loop feedback control mechanisms tackle this issue by degrading the network services gracefully to an acceptable level and then stabilizing the network in real-time (less than 50 seconds). Employing these feedback mechanisms also provide the ability to automatically configure the settings such that the QoS metrics of the network is consistent with those specified in the service level agreements.

Call number: LD2668 .T4 1985 S648 / Master of Science

Thesis (MScEng (Electrical and Electronic Engineering))--University of Stellenbosch, 2005. / Position calculation of mobile objects has challenged engineers and designers for years and is still continuing to do so. There are many solutions available today. Probably the best known and most widely used outdoor system today is the Global Positioning System (GPS). There are very little systems available for indoor use. An absolute positioning system was developed for this thesis. It uses a combination of ultrasonic and Radio Frequency (RF) communications to calculate a position fix in doors. Radar techniques were used to ensure robustness and reliability even in noisy environments. A small mobile robot was designed and built to test and illustrate the use of the system.

Type 2 diabetes has reached epidemic proportions worldwide. The resulting increase in chronic and costly diabetes related complications has potentially catastrophic implications for healthcare systems, and economics and societies as a whole. One of the key pathological factors leading to type 2 diabetes is insulin resistance (IR), which is the reduced or impaired ability of the body to make use of available insulin to maintain safe glucose concentrations in the bloodstream. It is essential to understand the physiology of glucose and insulin when investigating the underlying factors contributing to chronic diseases such as diabetes and cardiovascular disease. For many years, clinicians and researchers have been working to develop and use model-based methods to increase understanding and aid therapeutic decision support. However, the majority of practicable tests cannot yield more than basic metrics that allow only a threshold-based assessment of the underlying disorder. This thesis gives an overview on several dynamic model-based methodologies with different clinical applications in assessing glycaemia via measuring effects of treatment or medication on insulin sensitivity. Other tests are clinically focused, designed to screen populations and diagnose or detect the risk of developing diabetes. Thus, it is very important to observe sensitivity metrics in various clinical and research settings. Interstitial insulin kinetics and their influence on model-based insulin sensitivity observation was analysed using data from the clinical pilot study of the dynamic insulin sensitivity and secretion (DISST) test and the glucose-insulin PK-PD models. From these inputs, a model of interstitial insulin dose-response that best links insulin action in plasma to response in blood glucose levels was developed. The critical parameters influencing interstitial insulin pharmacokinetics (PKs) are saturation in insulin receptor binding (αG) and the plasma-interstitium diffusion rate (nI). Population values for these parameters are found to be [αG, nI]=[0.05,0.055]. Critically ill patients are regularly fed via constant enteral (EN) nutrition infusions. The impact of incretin effects on endogenous insulin secretion in this cohort remains unclear. It is hypothesised that the identified SI would decrease during interruptions of EN and would increase when EN is resumed, where, for short periods around transition, the true patient SI would be assumed constant. The model-based analysis was able to elucidate incretin effects by tracking the identified model-based insulin sensitivity (SI) in a cohort of critically ill patients. Thus, changes in model-based SI given the fixed assumed endogenous secretion by the model would support the presence of an EN-related incretin effect in the population of non-diabetic, critically ill patients studied. The PD feedback-control model of Uen was designed to investigate endogenous insulin secretion amongst subjects with different metabolic states and levels of insulin resistance. The underlying effects that influence insulin secretion i.e. incretin effects were also defined by tracking the control model gain/response and the identified insulin sensitivity (SI) using intravenous (IV) bolus and oral glucose responses of insulin sensitivity tests. This new PD control model allowed the characterisation of both static (basal) and dynamic insulin responses, which defined the pancreatic β-cell glucose sensitivity parameters. However, incretin effects were unobserved during oral glucose responses as the PD control gains failed to simulate the true endogenous insulin secretion due to potentially inaccurate glucose appearance rates and low data resolution of glucose concentrations. The net effect of haemodialysis (HD) treatment on glycaemic regulation and insulin sensitivity in a critically ill cohort was investigated. It was hypothesized that the observed SI would decrease during HD due to enhanced insulin clearance compared to the model, and would be recaptured again when HD is stopped. The changes in model-based SI metric at HD transitions in a cohort of critically ill patients were evaluated. Significant changes of -29% in model-based SI was observed during HD therapy. However, there were insignificant changes when HD treatment was ended. Thus, the changes in model-based SI would thus offer a unique observation on insulin kinetics and action in this population of critically ill patients with ARF that would better inform metabolic care.

On considère un écoulement compressible bidimensionnel, autour d'une cavité ouverte. Des d'instabilité, auto-entretenues par l'effet de rétroaction de l'écrasement de la couche de cisaillement sur le bord aval de la cavité, génèrent des émissions acoustiques qu'il faut réduire. Des simulations numériques directes (DNS) permettent d'obtenir, avec ou sans actionnement, un modèle précis de l'écoulement. A partir des champs issus de la simulation, des décompositions orthogonales de modes propres (POD) sont proposées pour bâtir, par projection de Galerkin sur les équations isentropiques, des modèles d'ordre réduit non linéaires en prenant en compte l'actionnement (le contrôle). Pour éviter la divergence temporelle, les coefficients du système dynamique non forcé sont calibrés par diverses approches originales dont une basée sur la sensiblité modale. A partir du système dynamique forcé par un actionnement multifréquentiel (présent aussi dans les DNS), un contrôle en boucle fermée linéaire quadratique gaussien est proposé sur un système linéarisé. La reconstruction de l'état est basée sur une estimation stochastique linéaire sur 6 points de pression. Le contrôle optimal obtenu s'avère être périodique à la fréquence du second mode de Rossiter, qui est exactement celles des instabilits auto-entretenues dans la cavité. Par introduction de ce contrôle dans les simulations numériques directes, nous avons obtenu une réduction du bruit (faible) sur la fréquence du contrôle. / We consider a two dimensional compressible flow around an open cavity. The Flow around a cavity is characterised by a self-sustained mechanism in which the shear layer impinges on the downstream edge of the cavity resulting in an acoustic feedback mechanism which must be reduced. Direct Numerical Simulations (DNS) of the flow at a representative Reynolds number has been carried to obtain pressure and velocity fields, both for the case of unactuated and multi frequency actuation. These fields are then used to extract energy ranked coherent structures also called as the Proper Orthogonal Decomposition (POD) modes. A Reduced Order Model is constructed by a Galerkin projections of the isentropic compressible equations. The model is then extended to include the effect of control. To avoid the divergence of the model while integrating in time various calibration techniques has been utillized. A new method of calibration which minimizes a linear functional of error, based on modal sensitivity is proposed. The calibrated low order model is used to design a feedback control of the Linear Quadratic Gaussian (LQG) type, coupled with an observer. For the experimental implementation of the controller, a state estimate based on the observed pressure measurements at 6 different locations, is obtained through a Linear Stochastic Estimation (LSE). The optimal control obtained is periodic with a frequency corresponding to the second Rossiter mode of the cavity. Finally the control obtained is introduced into the DNS to obtain a decrease in spectra of the cavity acoustic mode.

This work tries to remedy the practical part of teaching development of software for real-time systems. It does so by creation of a platform on which students can practically learn aspects of development of software for real-time systems. % (feedback control, low level programming). The resulting platform consists of a plant, a visualizer and a controller. The plant represents an industrial machine, the visualizer displays the current state of the plant. The controller drives the plant. Students learn by developing a program for the controller. The resulting platform is realized as a hardware-in-the-loop simulation -- the controller's processor and devices are real hardware, and the plant is a simulated device. The platform has a low cost, low space requirements and it is not easily breakable. Powered by TCPDF (www.tcpdf.org)

A dissertation submitted to the Faculty of Engineering, University of the Witwatersrand, .Johannesburg, in fulfilment of the requirements for the degree of Master of Science in Engineering. / As industrial plants become more complex. there is a growing need for new approaches to control and supervision. This research investigates the issues involved in applying Artificial Intelligence (AI) techniques in the real-world engineering problem of process control supervision. Current AI theory is examined and some techniques modified to design a general-purpose, reactive planner. The planner forms the basis of a supervisory control system. The system is implemented and interfaced with an existing Laboratory plant, so that its performance can he tested and evaluated by comparing with a conventional feedback controller This real life testing necessitates explicit treatment of issues such as data: sampling. situation assessment and CPU scheduling. The case study shows that by combining AI techniques with conventional control, a system can be built which displays superior performance under normal operating conditions and which can deal with abnormal conditions such as equipment failures. / Andrew Chakane 2018

Computação em nuvem refere-se a um modelo de disponibilização de recursos computacionais no qual a infraestrutura de software e hardware é ofertada como um serviço, e vem se estabelecendo como um paradigma de sucesso graças a versatilidade e ao custo-efetividade envolvidos nesse modelo de negócio, possibilitando o compartilhamento de um conjunto de recursos físicos entre diferentes usuários e aplicações. Com o advento da computação em nuvem e a possibilidade de elasticidade dos recursos computacionais virtualizados, a alocação dinâmica de recursos vem ganhando destaque, e com ela as questões referentes ao estabelecimento de contratos e de de qualidade de serviço. Historicamente, as pesquisas em QoS concentram-se na solução de problemas que envolvem duas entidades: usuários e servidores. Entretanto, em ambientes de nuvem, uma terceira entidade passa a fazer parte dessa interação, o consumidor de serviços em nuvem, que usa a infraestrutura para disponibilizar algum tipo de serviço aos usuários finais e que tem recebido pouca atenção das pesquisa até o momento, principalmente no que tange ao desenvolvimento de mecanismos automáticos para a alocação dinâmica de recursos sob variação de demanda. Este trabalho consiste na proposta de uma arquitetura de gerenciamento adaptativo de recursos sob a perspectiva do modelo de negócio envolvendo três entidades, focada na eficiência do consumidor. O trabalho inspira-se em técnicas de controle realimentado para encontrar soluções adaptativas aos problemas de alocação dinâmica de recursos, resultando em uma arquitetura de broker de consumidor, um respectivo protótipo e um método de projeto de controle para sistemas computacionais dessa natureza / CLoud computing refers to a computer resource deployment model in which software and hardware infrastructure are offered as a service. Cloud computing has become a successful paradigm due to the versatility and cost-effectiveness involved in that business model, making it possible to share a cluster of physical resources between several users and applications. With the advent of cloud computing and the computer elastic resource, dynamic allocation of virtualized resources is becoming more prominent, and along with it, the issues concerning the establishment of quality of service parameters. Historically, research on QoS has focused on solutions for problems involving two entities: users and servers. However, in cloud environments, a third party becomes part of this interaction, the cloud consumer, that uses the infrastructure to provide some kind of service to endusers, and which has received fewer attention, especially regarding the development of autonomic mechanisms for dynamic resource allocation under time-varying demand. This work aims at the development of an architecture for dynamic adaptive resource allocation involving three entities, focused on consumer revenue. The research outcome is a consumer broker architecture based on feedback control, a respective architecture prototype and a computer system feedback control methodology which may be applied in this class of problems