Model-based Self-awareness Patterns for Autonomy

Los sistemas técnicos son cada vez más complejos, incorporan funciones más avanzadas, están más integrados con otros sistemas y trabajan en entornos menos controlados. Todo esto supone unas condiciones más exigentes y con mayor incertidumbre para los sistemas de control, a los que además se demanda un comportamiento más autónomo y fiable. La adaptabilidad de manera autónoma es un reto para tecnologías de control actualmente. El proyecto de investigación ASys propone abordarlo trasladando la responsabilidad de la capacidad de adaptación del sistema de los ingenieros en tiempo de diseño al propio sistema en operación. Esta tesis pretende avanzar en la formulación y materialización técnica de los principios de ASys de cognición y auto-consciencia basadas en modelos y autogestión de los sistemas en tiempo de operación para una autonomía robusta. Para ello el trabajo se ha centrado en la capacidad de auto-conciencia, inspirada en los sistemas biológicos, y se ha explorado la posibilidad de integrarla en la arquitectura de los sistemas de control. Además de la auto-consciencia, se han explorado otros temas relevantes: modelado funcional, modelado de software, tecnología de los patrones, tecnología de componentes, tolerancia a fallos. Se ha analizado el estado de la técnica en los ámbitos pertinentes para las cuestiones de la auto-consciencia y la adaptabilidad en sistemas técnicos: arquitecturas cognitivas, control tolerante a fallos, y arquitecturas software dinámicas y computación autonómica. El marco teórico de ASys existente de sistemas autónomos cognitivos ha sido adaptado para servir de base para este análisis de autoconsciencia y adaptación y para dar sustento conceptual al posterior desarrollo de la solución. La tesis propone una solución general de diseño para la construcción de sistemas autónomos auto-conscientes. La idea central es la integración de un meta-controlador en la arquitectura de control del sistema autónomo, capaz de percibir la estado funcional del sistema de control y, si es necesario, reconfigurarlo en tiempo de operación. Esta solución de metacontrol se ha formalizado en cuatro patrones de diseño: i) el Patrón Metacontrol, que define la integración de un subsistema de metacontrol, responsable de controlar al propio sistema de control a través de la interfaz proporcionada por su plataforma de componentes, ii) el patrón Bucle de Control Epistémico, que define un bucle de control cognitivo basado en el modelos y que se puede aplicar al diseño del metacontrol, iii) el patrón de Reflexión basada en Modelo Profundo propone una solución para construir el modelo ejecutable utilizado por el meta-controlador mediante una transformación de modelo a modelo a partir del modelo de ingeniería del sistema, y, finalmente, iv) el Patrón Metacontrol Funcional, que estructura el meta-controlador en dos bucles, uno para el control de la configuración de los componentes del sistema de control, y otro sobre éste, controlando las funciones que realiza dicha configuración de componentes; de esta manera las consideraciones funcionales y estructurales se desacoplan. La Arquitectura OM y el metamodelo TOMASys son las piezas centrales del marco arquitectónico desarrollado para materializar la solución compuesta de los patrones anteriores. El metamodelo TOMASys ha sido desarrollado para la representación de la estructura y su relación con los requisitos funcionales de cualquier sistema autónomo. La Arquitectura OM es un patrón de referencia para la construcción de una metacontrolador integrando los patrones de diseño propuestos. Este meta-controlador se puede integrar en la arquitectura de cualquier sistema control basado en componentes. El elemento clave de su funcionamiento es un modelo TOMASys del sistema decontrol, que el meta-controlador usa para monitorizarlo y calcular las acciones de reconfiguración necesarias para adaptarlo a las circunstancias en cada momento. Un proceso de ingeniería, complementado con otros recursos, ha sido elaborado para guiar la aplicación del marco arquitectónico OM. Dicho Proceso de Ingeniería OM define la metodología a seguir para construir el subsistema de metacontrol para un sistema autónomo a partir del modelo funcional del mismo. La librería OMJava proporciona una implementación del meta-controlador OM que se puede integrar en el control de cualquier sistema autónomo, independientemente del dominio de la aplicación o de su tecnología de implementación. Para concluir, la solución completa ha sido validada con el desarrollo de un robot móvil autónomo que incorpora un meta-controlador con la Arquitectura OM. Las propiedades de auto-consciencia y adaptación proporcionadas por el meta-controlador han sido validadas en diferentes escenarios de operación del robot, en los que el sistema era capaz de sobreponerse a fallos en el sistema de control mediante reconfiguraciones orquestadas por el metacontrolador. ABSTRACT Technical systems are becoming more complex, they incorporate more advanced functionalities, they are more integrated with other systems and they are deployed in less controlled environments. All this supposes a more demanding and uncertain scenario for control systems, which are also required to be more autonomous and dependable. Autonomous adaptivity is a current challenge for extant control technologies. The ASys research project proposes to address it by moving the responsibility for adaptivity from the engineers at design time to the system at run-time. This thesis has intended to advance in the formulation and technical reification of ASys principles of model-based self-cognition and having systems self-handle at runtime for robust autonomy. For that it has focused on the biologically inspired capability of self-awareness, and explored the possibilities to embed it into the very architecture of control systems. Besides self-awareness, other themes related to the envisioned solution have been explored: functional modeling, software modeling, patterns technology, components technology, fault tolerance. The state of the art in fields relevant for the issues of self-awareness and adaptivity has been analysed: cognitive architectures, fault-tolerant control, and software architectural reflection and autonomic computing. The extant and evolving ASys Theoretical Framework for cognitive autonomous systems has been adapted to provide a basement for this selfhood-centred analysis and to conceptually support the subsequent development of our solution. The thesis proposes a general design solution for building self-aware autonomous systems. Its central idea is the integration of a metacontroller in the control architecture of the autonomous system, capable of perceiving the functional state of the control system and reconfiguring it if necessary at run-time. This metacontrol solution has been formalised into four design patterns: i) the Metacontrol Pattern, which defines the integration of a metacontrol subsystem, controlling the domain control system through an interface provided by its implementation component platform, ii) the Epistemic Control Loop pattern, which defines a modelbased cognitive control loop that can be applied to the design of such a metacontroller, iii) the Deep Model Reflection pattern proposes a solution to produce the online executable model used by the metacontroller by model-to-model transformation from the engineering model, and, finally, iv) the Functional Metacontrol pattern, which proposes to structure the metacontroller in two loops, one for controlling the configuration of components of the controller, and another one on top of the former, controlling the functions being realised by that configuration; this way the functional and structural concerns become decoupled. The OM Architecture and the TOMASys metamodel are the core pieces of the architectural framework developed to reify this patterned solution. The TOMASys metamodel has been developed for representing the structure and its relation to the functional requirements of any autonomous system. The OM architecture is a blueprint for building a metacontroller according to the patterns. This metacontroller can be integrated on top of any component-based control architecture. At the core of its operation lies a TOMASys model of the control system. An engineering process and accompanying assets have been constructed to complete and exploit the architectural framework. The OM Engineering Process defines the process to follow to develop the metacontrol subsystem from the functional model of the controller of the autonomous system. The OMJava library provides a domain and application-independent implementation of an OM Metacontroller than can be used in the implementation phase of OMEP. Finally, the complete solution has been validated in the development of an autonomous mobile robot that incorporates an OM metacontroller. The functional selfawareness and adaptivity properties achieved thanks to the metacontrol system have been validated in different scenarios. In these scenarios the robot was able to overcome failures in the control system thanks to reconfigurations performed by the metacontroller.

Model Based On-Line Energy Prediction System for Semi-Autonomous Mobile Robots

Maximizing energy autonomy is a consistent challenge when deploying mobile robots in ionizing radiation or other hazardous environments. Having a reliable robot system is essential for successful execution of missions and to avoid manual recovery of the robots in environments that are harmful to human beings. For deployment of robots missions at short notice, the ability to know beforehand the energy required for performing the task is essential. This paper presents a on-line method for predicting energy requirements based on the pre-determined power models for a mobile robot. A small mobile robot, Khepera III is used for the experimental study and the results are promising with high prediction accuracy. The applications of the energy prediction models in energy optimization and simulations are also discussed along with examples of significant energy savings.

Integrating structural and input design of a 2-DOF high-speed parallel manipulator: A flexible model-based approach

This paper discusses the integrated design of parallel manipulators, which exhibit varying dynamics. This characteristic affects the machine stability and performance. The design methodology consists of four main steps: (i) the system modeling using flexible multibody technique, (ii) the synthesis of reduced-order models suitable for control design, (iii) the systematic flexible model-based input signal design, and (iv) the evaluation of some possible machine designs. The novelty in this methodology is to take structural flexibilities into consideration during the input signal design; therefore, enhancing the standard design process which mainly considers rigid bodies dynamics. The potential of the proposed strategy is exploited for the design evaluation of a two degree-of-freedom high-speed parallel manipulator. The results are experimentally validated. (C) 2010 Elsevier Ltd. All rights reserved.

New prediction methods for CO2 evaporation inside tubes: Part II - An updated general flow boiling heat transfer model based on flow patterns

Corresponding to the updated flow pattern map presented in Part I of this study, an updated general flow pattern based flow boiling heat transfer model was developed for CO2 using the Cheng-Ribatski-Wojtan-Thome [L. Cheng, G. Ribatski, L. Wojtan, J.R. Thome, New flow boiling heat transfer model and flow pattern map for carbon dioxide evaporating inside horizontal tubes, Int. J. Heat Mass Transfer 49 (2006) 4082-4094; L. Cheng, G. Ribatski, L. Wojtan, J.R. Thome, Erratum to: ""New flow boiling heat transfer model and flow pattern map for carbon dioxide evaporating inside tubes"" [Heat Mass Transfer 49 (21-22) (2006) 4082-4094], Int. J. Heat Mass Transfer 50 (2007) 391] flow boiling heat transfer model as the starting basis. The flow boiling heat transfer correlation in the dryout region was updated. In addition, a new mist flow heat transfer correlation for CO2 was developed based on the CO2 data and a heat transfer method for bubbly flow was proposed for completeness sake. The updated general flow boiling heat transfer model for CO2 covers all flow regimes and is applicable to a wider range of conditions for horizontal tubes: tube diameters from 0.6 to 10 mm, mass velocities from 50 to 1500 kg/m(2) s, heat fluxes from 1.8 to 46 kW/m(2) and saturation temperatures from -28 to 25 degrees C (reduced pressures from 0.21 to 0.87). The updated general flow boiling heat transfer model was compared to a new experimental database which contains 1124 data points (790 more than that in the previous model [Cheng et al., 2006, 2007]) in this study. Good agreement between the predicted and experimental data was found in general with 71.4% of the entire database and 83.2% of the database without the dryout and mist flow data predicted within +/-30%. However, the predictions for the dryout and mist flow regions were less satisfactory due to the limited number of data points, the higher inaccuracy in such data, scatter in some data sets ranging up to 40%, significant discrepancies from one experimental study to another and the difficulties associated with predicting the inception and completion of dryout around the perimeter of the horizontal tubes. (C) 2007 Elsevier Ltd. All rights reserved.

Melphalan dosing regimens for management of recurrent melanoma by isolated limb perfusion: Application of a physiological pharmacokinetic model based on melphalan distribution in the isolated perfused rat hindlimb

The optimal dosing schedule for melphalan therapy of recurrent malignant melanoma in isolated limb perfusions has been examined using a physiological pharmacokinetic model with data from isolated rat hindlimb perfusions (IRHP), The study included a comparison of melphalan distribution in IRHP under hyperthermia and normothermia conditions. Rat hindlimbs were perfused with Krebs-Henseleit buffer containing 4.7% bovine serum albumin at 37 or 41.5 degrees C at a flow rate of 4 ml/min. Concentrations of melphalan in perfusate and tissues were determined by high performance liquid chromatography with fluorescence detection, The concentration of melphalan in perfusate and tissues was linearly related to the input concentration. The rate and amount of melphalan uptake into the different tissues was higher at 41.5 degrees C than at 37 degrees C. A physiological pharmacokinetic model was validated from the tissue and perfusate time course of melphalan after melphalan perfusion. Application of the model involved the amount of melphalan exposure in the muscle, skin and fat in a recirculation system was related to the method of melphalan administration: single bolus > divided bolus > infusion, The peak concentration of melphalan in the perfusate was also related to the method of administration in the same order, Infusing the total dose of melphalan over 20 min during a 60 min perfusion optimized the exposure of tissues to melphalan whilst minimizing the peak perfusate concentration of melphalan. It is suggested that this method of melphalan administration may be preferable to other methods in terms of optimizing the efficacy of melphalan whilst minimizing the limb toxicity associated with its use in isolated limb perfusion.

Diagnostic model based on Raman spectra of normal, hyperplasia and prostate adenocarcinoma tissues in vitro

This study evaluated the use of Raman spectroscopy to identify the spectral differences between normal (N), benign hyperplasia (BPH) and adenocarcinoma (CaP) in fragments of prostate biopsies in vitro with the aim of developing a spectral diagnostic model for tissue classification. A dispersive Raman spectrometer was used with 830 nm wavelength and 80 mW excitation. Following Raman data collection and tissue histopathology (48 fragments diagnosed as N, 43 as BPH and 14 as CaP), two diagnostic models were developed in order to extract diagnostic information: the first using PCA and Mahalanobis analysis techniques and the second one a simplified biochemical model based on spectral features of cholesterol, collagen, smooth muscle cell and adipocyte. Spectral differences between N, BPH and CaP tissues, were observed mainly in the Raman bands associated with proteins, lipids, nucleic and amino acids. The PCA diagnostic model showed a sensitivity and specificity of 100%, which indicates the ability of PCA and Mahalanobis distance techniques to classify tissue changes in vitro. Also, it was found that the relative amount of collagen decreased while the amount of cholesterol and adipocyte increased with severity of the disease. Smooth muscle cell increased in BPH tissue. These characteristics were used for diagnostic purposes.

Classification model based on Raman spectra of selected morphological and biochemical tissue constituents for identification of atherosclerosis in human coronary arteries

This study presents the results of Raman spectroscopy applied to the classification of arterial tissue based on a simplified model using basal morphological and biochemical information extracted from the Raman spectra of arteries. The Raman spectrograph uses an 830-nm diode laser, imaging spectrograph, and a CCD camera. A total of 111 Raman spectra from arterial fragments were used to develop the model, and those spectra were compared to the spectra of collagen, fat cells, smooth muscle cells, calcification, and cholesterol in a linear fit model. Non-atherosclerotic (NA), fatty and fibrous-fatty atherosclerotic plaques (A) and calcified (C) arteries exhibited different spectral signatures related to different morphological structures presented in each tissue type. Discriminant analysis based on Mahalanobis distance was employed to classify the tissue type with respect to the relative intensity of each compound. This model was subsequently tested prospectively in a set of 55 spectra. The simplified diagnostic model showed that cholesterol, collagen, and adipocytes were the tissue constituents that gave the best classification capability and that those changes were correlated to histopathology. The simplified model, using spectra obtained from a few tissue morphological and biochemical constituents, showed feasibility by using a small amount of variables, easily extracted from gross samples.

Model based procedure for scale-up of wet overflow ball mills Part 1: Outline of the methodology

Bond's method for ball mill scale-up only gives the mill power draw for a given duty. This method is incompatible with computer modelling and simulation techniques. It might not be applicable for the design of fine grinding ball mills and ball mills preceded by autogenous and semi-autogenous grinding mills. Model-based ball mill scale-up methods have not been validated using a wide range of full-scale circuit data. Their accuracy is therefore questionable. Some of these methods also need expensive pilot testing. A new ball mill scale-up procedure is developed which does not have these limitations. This procedure uses data from two laboratory tests to determine the parameters of a ball mill model. A set of scale-up criteria then scales-up these parameters. The procedure uses the scaled-up parameters to simulate the steady state performance of full-scale mill circuits. At the end of the simulation, the scale-up procedure gives the size distribution, the volumetric flowrate and the mass flowrate of all the streams in the circuit, and the mill power draw.

Model-based procedure for scale-up of wet, overflow ball mills - Part II: Worked example

A new ball mill scale-up procedure is developed which uses laboratory data to predict the performance of MI-scale ball mill circuits. This procedure contains two laboratory tests. These laboratory tests give the data for the determination of the parameters of a ball mill model. A set of scale-up criteria then scales-up these parameters. The procedure uses the scaled-up parameters to simulate the steady state performance of the full-scale mill circuit. At the end of the simulation, the scale-up procedure gives the size distribution, the volumetric flowrate and the mass flowrate of all the streams in the circuit, and the mill power draw. A worked example shows how the new ball mill scale-up procedure is executed. This worked example uses laboratory data to predict the performance of a full-scale re-grind mill circuit. This circuit consists of a ball mill in closed circuit with hydrocyclones. The MI-scale ball mill has a diameter (inside liners) of 1.85m. The scale-up procedure shows that the full-scale circuit produces a product (hydrocyclone overflow) that has an 80% passing size of 80 mum. The circuit has a recirculating load of 173%. The calculated power draw of the full-scale mill is 92kW (C) 2001 Elsevier Science Ltd. All rights reserved.

Model-based procedure for scale-up of wet, overflow ball mills - Part III: Validation and discussion

A new ball mill scale-up procedure is developed. This procedure has been validated using seven sets of Ml-scale ball mil data. The largest ball mills in these data have diameters (inside liners) of 6.58m. The procedure can predict the 80% passing size of the circuit product to within +/-6% of the measured value, with a precision of +/-11% (one standard deviation); the re-circulating load to within +/-33% of the mass-balanced value (this error margin is within the uncertainty associated with the determination of the re-circulating load); and the mill power to within +/-5% of the measured value. This procedure is applicable for the design of ball mills which are preceded by autogenous (AG) mills, semi-autogenous (SAG) mills, crushers and flotation circuits. The new procedure is more precise and more accurate than Bond's method for ball mill scale-up. This procedure contains no efficiency correction which relates to the mill diameter. This suggests that, within the range of mill diameter studied, milling efficiency does not vary with mill diameter. This is in contrast with Bond's equation-Bond claimed that milling efficiency increases with mill diameter. (C) 2001 Elsevier Science Ltd. All rights reserved.

A mixture model-based approach to the clustering of microarray expression data

Motivation: This paper introduces the software EMMIX-GENE that has been developed for the specific purpose of a model-based approach to the clustering of microarray expression data, in particular, of tissue samples on a very large number of genes. The latter is a nonstandard problem in parametric cluster analysis because the dimension of the feature space (the number of genes) is typically much greater than the number of tissues. A feasible approach is provided by first selecting a subset of the genes relevant for the clustering of the tissue samples by fitting mixtures of t distributions to rank the genes in order of increasing size of the likelihood ratio statistic for the test of one versus two components in the mixture model. The imposition of a threshold on the likelihood ratio statistic used in conjunction with a threshold on the size of a cluster allows the selection of a relevant set of genes. However, even this reduced set of genes will usually be too large for a normal mixture model to be fitted directly to the tissues, and so the use of mixtures of factor analyzers is exploited to reduce effectively the dimension of the feature space of genes. Results: The usefulness of the EMMIX-GENE approach for the clustering of tissue samples is demonstrated on two well-known data sets on colon and leukaemia tissues. For both data sets, relevant subsets of the genes are able to be selected that reveal interesting clusterings of the tissues that are either consistent with the external classification of the tissues or with background and biological knowledge of these sets.

Model-based clustering in gene expression microarrays: An application to breast cancer data

In microarray studies, the application of clustering techniques is often used to derive meaningful insights into the data. In the past, hierarchical methods have been the primary clustering tool employed to perform this task. The hierarchical algorithms have been mainly applied heuristically to these cluster analysis problems. Further, a major limitation of these methods is their inability to determine the number of clusters. Thus there is a need for a model-based approach to these. clustering problems. To this end, McLachlan et al. [7] developed a mixture model-based algorithm (EMMIX-GENE) for the clustering of tissue samples. To further investigate the EMMIX-GENE procedure as a model-based -approach, we present a case study involving the application of EMMIX-GENE to the breast cancer data as studied recently in van 't Veer et al. [10]. Our analysis considers the problem of clustering the tissue samples on the basis of the genes which is a non-standard problem because the number of genes greatly exceed the number of tissue samples. We demonstrate how EMMIX-GENE can be useful in reducing the initial set of genes down to a more computationally manageable size. The results from this analysis also emphasise the difficulty associated with the task of separating two tissue groups on the basis of a particular subset of genes. These results also shed light on why supervised methods have such a high misallocation error rate for the breast cancer data.

Model-based analysis of anaerobic acetate uptake by a mixed culture of polyphosphate-accumulating and glycogen-accumulating organisms

An increasing number of studies shows that the glycogen-accumulating organisms (GAOs) can survive and may indeed proliferate under the alternating anaerobic/aerobic conditions found in EBPR systems, thus forming a strong competitor of the polyphosphate-accumulating organisms (PAOs). Understanding their behaviors in a mixed PAO and GAO culture under various operational conditions is essential for developing operating strategies that disadvantage the growth of this group of unwanted organisms. A model-based data analysis method is developed in this paper for the study of the anaerobic PAO and GAO activities in a mixed PAO and GAO culture. The method primarily makes use of the hydrogen ion production rate and the carbon dioxide transfer rate resulting from the acetate uptake processes by PAOs and GAOs, measured with a recently developed titration and off-gas analysis (TOGA) sensor. The method is demonstrated using the data from a laboratory-scale sequencing batch reactor (SBR) operated under alternating anaerobic and aerobic conditions. The data analysis using the proposed method strongly indicates a coexistence of PAOs and GAOs in the system, which was independently confirmed by fluorescent in situ hybridization (FISH) measurement. The model-based analysis also allowed the identification of the respective acetate uptake rates by PAOs and GAOs, along with a number of kinetic and stoichiometric parameters involved in the PAO and GAO models. The excellent fit between the model predictions and the experimental data not involved in parameter identification shows that the parameter values found are reliable and accurate. It also demonstrates that the current anaerobic PAO and GAO models are able to accurately characterize the PAO/GAO mixed culture obtained in this study. This is of major importance as no pure culture of either PAOs or GAOs has been reported to date, and hence the current PAO and GAO models were developed for the interpretation of experimental results of mixed cultures. The proposed method is readily applicable for detailed investigations of the competition between PAOs and GAOs in enriched cultures. However, the fermentation of organic substrates carried out by ordinary heterotrophs needs to be accounted for when the method is applied to the study of PAO and GAO competition in full-scale sludges. (C) 2003 Wiley Periodicals, Inc.