The development of the nonlinear version of the Continuous-time Generalised Predictive Control (NCGPC) is presented. Unlike the linear version, the nonlinear version is developed in state-space form and shown to include Nonlinear Generalised Minimum Variance (NGMV), and a new algorithm, Nonlinear Predictive Generalised Minimum Variance (NPGMV), as special cases. Through simulations, it is demonstrated that NCGPC can deal with nonlinear systems whose relative degree is not well defined and nonlinear systems with unstable zero dynamics. Geometric approaches, such as exact linearisation, are shown to be included in the NCGPC as special cases.

Although human sustained control movements are continuous in nature there is still controversy on the mechanisms underlying such physiological systems. A popular topic of debate is whether human motor control mechanisms could be modelled as engineering control systems, and if so, what control algorithm is most appropriate. Since the early years of modelling sustained control tasks in human motor control the servomechanism has been an adequate model to describe human tracking tasks. Another continuous-time system model that is often used to model sustained control tasks is the predictive controller which is based on internal models and includes prediction and optimisation. On the other hand, studies have suggested intermittent behaviour of the ``human controller'' in sustained motor control tasks. This thesis investigated whether intermittent control is a suitable approach to describe sustained human motor control. It was investigated how well an intermittent control system model could approximate both the deterministic and non-deterministic parts of experimental data, from a visual-manual compensatory tracking task. Finally, a preliminary study was conducted to explore issues associated with the practical implementation of the intermittent control model. To fit the deterministic part of experimental data, a frequency domain identification method was used. Identification results obtained with an intermittent controller were compared against the results using continuous-time non-predictive and predictive controllers. The results show that the identified frequency response functions of the intermittent control model not only fit the frequency response functions derived from the experimental data well, but most importantly resulted in identified controller parameters which are similar to those identified using a predictive controller, and whose parameter values appear to be physiologically meaningful. A novel way to explain human variability, as represented by the non-deterministic part of the experimental data (the \emph{remnant}), was developed, based on an intermittent control model with variable intermittent interval. This model was compared against the established paradigm, in which variability is explained by a predictive controller with added noise, either signal dependent control signal noise, or observation noise. The study has shown that the intermittent controller with a variable intermittent interval could model the non-deterministic experimental data as well as the predictive controller model with added noise. This provides a new explanation for the source of remnant in human control as inherent to the controller structure, rather than as a noise signal, and enables a new interpretation for the physiological basis for human variability. Finally, the theoretical intermittent control model was implemented in real-time in the context of the physiological control mechanism of human standing balance. An experimental method was developed to apply automatic artificial balance of an inverted pendulum in the context of human standing, via functions electrical stimulation control of the lower leg muscles of a healthy subject. The significance of this study is, firstly, that frequency domain identification was applied for the first time with intermittent control, and it could be shown that both intermittent and predictive control models can model deterministic experimental data from manual tracking tasks equally well. Secondly, for the first time the inherent variability, which is represented by the remnant signal, in human motor control tasks could be modelled as part of the structure of the intermittent controller rather than as an added noise model. Although, the experimental method to apply automatic artificial balance of an inverted pendulum in the context of human standing was not successful, the intermittent controller was implemented for the first time in real-time and combined with electrical muscle stimulation to control a physiological mechanism.

No description available.

Covariance and Gramian matrices in control and systems theory and pattern recognition are studied in the context of reduction of dimensionality and hence complexity of large-scale systems. This is achieved by the removal of redundant or 'almost' redundant information contained in the covariance and Grarrdan matrices. The Karhunen-Loeve expansion (principal component analysis) and its extensions and the singular value decomposition of matrices provide the framework for the work presented in the thesis. The results given for linear dynamical systems are based on controllability and observability Gramians and some new developments in singular perturbational analysis are also presented.

Le travail de recherche initié dans cette thèse de doctorat s’inscrit dans le contexte du projet européen InTraDE (Intelligent Transportation for Dynamic Environment). Les aspects théoriques développés concernent la modélisation graphique de Système de systèmes (SdS). Ces derniers peuvent être modélisés sur plusieurs niveaux, allant du niveau microscopique au niveau macroscopique. Une approche à base d'hypergraphe est utilisée pour représenter les différents niveaux hiérarchiques d’un SdS ainsi que les propriétés fondamentales qui le définissent : l'indépendance, la coopération, la dispersion, et l'évolution des composants systèmes qui le constituent. La modélisation graphique proposée permet d’élaborer une stratégie de supervision en ligne en suivant les parcours ascendant et descendant de l'hypergraphe. Pour gérer le fonctionnement global des composants systèmes à un niveau de modélisation donné, un modèle graphique dynamique est extrait de l'hypergraphe. Ce dernier, permet d’appliquer une stratégie d'optimisation en ligne. Appliquée au système de transport intelligent, cette approche permet de déterminer à chaque instant le temps que mettra le véhicule autonome et intelligent pour atteindre sa cible lorsqu’il évolue dans un environnement évoluant en permanence. / This work has been realized in the framework of the European project InTraDE (Intelligent Transportation for Dynamic Environment). The main scientific contribution focuses on graphical modeling of System of Systems (SoS) in the engineering field. To model such systems, a graphical method based on hypergraph approach is proposed. Hypergraphs are used to describe the multiple hierarchical levels of a SoS and to describe its associated fundamental properties such as: the independence, cooperation, dispersion and the evolution of the component systems. Through this graphical modeling, online supervision strategies can be directly deduced. To manage the global operating of the component systems at a specific level, a dynamic graphical model is extracted from the hypergraph. This later allows the elaboration of an online optimization strategy. As a case study, an intelligent transportation system is considered. An online optimization algorithm currently updates, the time used for each intelligent autonomous vehicle to reach its target, in a permanently evolving environment.

No description available.

The use of multiple robots for exploration holds the promise of improved performance over single robot systems. To exploit effectively the advantage of having several robots, the robots must be co-ordinated which requires communication. Previous research relies on a fixed communication network topology, a single lead explorer, and flat communication. This thesis presents a novel architecture to keep a group of robots as a single connected and adaptable communication network to explore and map the environment. This architecture, BERODE (BEhavioural ROle DEcentralized), aims to be robust, efficient and scalable to large numbers of robots. The network is adaptable, the number of explorers variable, and communications hierarchical (local/global). The network is kept connected by an MST (Minimum Spanning Tree) control network, a subnetwork containing only the minimum necessary links to be a fully connected network. As the robots explore, the MST control network is updated either partially (local network) or globally to improve signal quality. The local network for a robot is formed by the robots that are within a certain retransmission distance in the MST control network. BERODE implements a hierarchic approach to distributing information to improve scalability with respect to the number of robots. The robots share information at two levels: frequently within their local network and less frequently to the entire robot network. The robots coordinate by assuming behaviours depending on their connections in the MST control network. The behavioural roles balance between the tasks of exploration and network maintenance where the Explorer role is the most focused on the exploration task. This improves efficiency by allowing varying number of robots to take the Explorer role depending on circumstances. The roles generate reactive plans that ensure the connectivity of the network. These plans are based on the imposition of heterogeneous virtual spring forces. Our simulations show that BERODE is more efficient, scalable and robust with respect to communications than the previous approaches that rely on fixed control networks. BERODE is more efficient because it required less time to build a complete map of the environment than the fixed control networks. BERODE is more scalable because it keeps the robots as a single connected network for more time than the fixed control networks. BERODE is more robust because it has a better success rate at finishing the exploration.

No description available.

The evaluation of risk within a scene is a new and emerging area of research. With the advent of smart enabled homes and the continued development and implementation of domestic robotics, the platform for automated risk assessment within the home is now a possibility. The aim of this thesis is to explore a subsection of the problems facing the detection and quantification of risk in a domestic setting. A Risk Estimation framework is introduced which provides a flexible and context aware platform from which measurable elements of risk can be combined to create a final risk score for a scene. To populate this framework, three elements of measurable risk are proposed and evaluated: Firstly, scene stability, assessing the location and stability of objects within an environment through the use of physics simulation techniques. Secondly, hazard feature analysis using two specifically designed novel feature descriptors (3D Voxel HOG and the Physics Behaviour Feature) which determine if the objects within a scene have dangerous or risky properties such as blades or points. Finally, environment interaction, which uses human behaviour simulation to predict human reactions to detected risks and highlight areas of a scene most likely to be visited. Additionally methodologies are introduced to support these concepts including: a simulation prediction framework which reduces the computational cost of physics simulation, a Robust Filter and Complex Adaboost which aim to improve the robustness and training times required for hazard feature classification models. The Human and Group Behaviour Evaluation framework is introduced to provide a platform from which simulation algorithms can be evaluated without the need for extensive ground truth data. Finally the 3D Risk Scenes (3DRS) dataset is introduced, creating a risk specific dataset for the evaluation of future domestic risk analysis methodologies.

The primary objective of this research is to develop a general modelling facility for modular pneumatic servo-drives. The component-oriented approach has been adopted as the modelling technique to provide the flexibility of modelling a wide variety of components and the segmentation of the non-linear system to less complex uncoupled component modules. A significant part of the research work has been devoted to identify a series of component modules of the single axis linear pneumatic servomechanism with standardised linking variables. The mathematical models have been implemented in a simulation software which produces time domain responses for design evaluation purposes. Alternative components for different servomechanism design were modelled as mutually exclusive modules which could be selected for assembly as if they were real physical entities. The philosophy of the approach was validated by tests on prototype servo-drives with matching components. Design analysis could be performed by simulating and comparing the performance of alternative system structures.