This work presents a novel approach to modeling, analysis and diagnosis of coupled mechatronical systems with partially autonomous behavior and asynchronous state transitions. The systems under consideration are assumed to have the following properties: The internal interactions are immeasurable but reliable and the measurements relevant for diagnosis are given as a sequence of events. Asynchronous networks of input/output automata (I/O-automata) are developed to cope with partial coupling between components and to reduce the computational complexity of the diagnostic algorithms. I/O-automata are used to model those components. Their measurable inputs and outputs are modeled as control signals. Interconnection signals are used to model the internal dependencies among the components. They are linked via an interaction block to one another. The criterion known from synchronous networks of I/O-automata is extended to ensure the well-posedness of this modeling formalism. To check for partially autonomous behavior, two types of autonomy are introduced and discussed: Structural autonomy and state-dependent autonomy. To carry out the diagnosis, three different information structures are investigated: Centralized, decentralized and partially coordinated. The centralized approach yields the ideal diagnostic result, but reduction of the computational complexity by using online composition is rather small. Further reduction of the computational complexity is accomplished by decentralized diagnosis. It yields only in the case of state-dependent autonomy a complete and sound diagnostic result. In general, the lack of soundness arises. Both, obtaining an ideal diagnostic result and reducing the computational complexity, is obtained by the partially coordinated diagnostic algorithm.

This book presents novel methods of fault-tolerant control theory in a discrete-event system framework. Nondeterministic input/output automata are used to model nominal and faulty technological systems. The main contributions are the following: Control design method for discrete-event systems Fault modeling technique for actuator, sensor and system internal faults and failures Off-line and on-line control reconfiguration based on trajectory re-planning and input/output adaptation. Two small size running examples are used to explain the developed methods. Experiments on a manufacturing cell demonstrate the application of these methods in a realistic environment. The state of the art is provided on methods for modeling, supervisory control and fault-tolerant control of discrete-event systems.

Estimation and Inference in Discrete Event Systems chooses a popular model for emerging automation systems—finite automata under partial observation—and focuses on a comprehensive study of the key problems of state estimation and event inference. The text includes treatment of current, delayed, and initial state estimation. Related applications for assessing and enforcing resiliency—fault detection and diagnosis—and security—privacy and opacity—properties are discussed, enabling the reader to apply these techniques in a variety of emerging applications, among them automated manufacturing processes, intelligent vehicle/highway systems, and autonomous vehicles. The book provides a systematic development of recursive algorithms for state estimation and event inference. The author also deals with the verification of pertinent properties such as: the ability to determine the exact state of a system, “detectability”; the ability to ensure that certain classes of faults can be detected/identified, “diagnosability”; and the ability to ensure that certain internal state variables of the system remain “hidden” from the outside world regardless of the type of activity that is taking place, “opacity”. This book allows students, researchers and practicing engineers alike to grasp basic aspects of state estimation in discrete event systems, aspects like distributivity and probabilistic inference, quickly and without having to master the entire breadth of models that are available in the literature.

This book constitutes the refereed proceedings of the 32nd International Conference on Applications and Theory of Petri Nets and Other Models of Concurrency, PETRI NETS 2011, held in Newcastle, UK, in June 2011. The 13 regular papers and 4 tool papers presented were carefully reviewed and selected from 49 submissions. The book also contains 3 full paper length invited talks. All current issues on research and development in the area of Petri nets and related models of concurrent systems are addressed.

Illustrated with real-life manufacturing examples, Formal Methods in Manufacturing provides state-of-the-art solutions to common problems in manufacturing systems. Assuming some knowledge of discrete event systems theory, the book first delivers a detailed introduction to the most important formalisms used for the modeling, analysis, and control of manufacturing systems (including Petri nets, automata, and max-plus algebra), explaining the advantages of each formal method. It then employs the different formalisms to solve specific problems taken from today’s industrial world, such as modeling and simulation, supervisory control (including deadlock prevention) in a distributed and/or decentralized environment, performance evaluation (including scheduling and optimization), fault diagnosis and diagnosability analysis, and reconfiguration. Containing chapters written by leading experts in their respective fields, Formal Methods in Manufacturing helps researchers and application engineers handle fundamental principles and deal with typical quality goals in the design and operation of manufacturing systems.

Discrete Event Systems: Diagnosis and Diagnosability addresses the problem of fault diagnosis of Discrete Event Systems (DESs). This book provides the basic techniques and approaches necessary for the design of an efficient fault diagnosis system for a wide range of modern engineering applications. This book classifies the different techniques and approaches according to several criteria such as: modeling tools (Automata, Petri nets, Templates) that is used to construct the model; the information (qualitative based on events occurrences and/or states outputs, quantitative based on signal processing, data analysis) that is needed to analyze and achieve the diagnosis; the decision structure (centralized, decentralized) that is required to achieve the diagnosis; as well as the complexity (polynomial, exponential) of the algorithm that is used to determine the set of faults that the proposed approach is able to diagnose as well as the delay time required for this diagnosis. The goal of this classification is to select the efficient method to achieve the fault diagnosis according to the application constraints. This book will include illustrated examples of the presented methods and techniques as well as a discussion on the application of these methods on several real-world problems.

This book describes how control of distributed systems can be advanced by an integration of control, communication, and computation. The global control objectives are met by judicious combinations of local and nonlocal observations taking advantage of various forms of communication exchanges between distributed controllers. Control architectures are considered according to increasing degrees of cooperation of local controllers: fully distributed or decentralized control, control with communication between controllers, coordination control, and multilevel control. The book covers also topics bridging computer science, communication, and control, like communication for control of networks, average consensus for distributed systems, and modeling and verification of discrete and of hybrid systems. Examples and case studies are introduced in the first part of the text and developed throughout the book. They include: control of underwater vehicles, automated-guided vehicles on a container terminal, control of a printer as a complex machine, and control of an electric power system. The book is composed of short essays each within eight pages, including suggestions and references for further research and reading. By reading the essays collected in the book Coordination Control of Distributed Systems, graduate students and post-docs will be introduced to the research frontiers in control of decentralized and of distributed systems. Control theorists and practitioners with backgrounds in electrical, mechanical, civil and aerospace engineering will find in the book information and inspiration to transfer to their fields of interest the state-of-art in coordination control.

Annotation This work presents a novel approach to modeling, analysis and diagnosis of coupled mechatronical systems with partially autonomous behavior and asynchronous state transitions. The systems under consideration are assumed to have the following properties: The internal interactions are immeasurable but reliable and the measurements relevant for diagnosis are given as a sequence of events. Asynchronous networks of input/output automata (I/O-automata) are developed to cope with partial coupling between components and to reduce the computational complexity of the diagnostic algorithms. I/O-automata are used to model those components. Their measurable inputs and outputs are modeled as control signals. Interconnection signals are used to model the internal dependencies among the components. They are linked via an interaction block to one another. The criterion known from synchronous networks of I/O-automata is extended to ensure the well-posedness of this modeling formalism. To check for partially autonomous behavior, two types of autonomy are introduced and discussed: Structural autonomy and state-dependent autonomy. To carry out the diagnosis, three different information structures are investigated: Centralized, decentralized and partially coordinated. The centralized approach yields the ideal diagnostic result, but reduction of the computational complexity by using online composition is rather small. Further reduction of the computational complexity is accomplished by decentralized diagnosis. It yields only in the case of state-dependent autonomy a complete and sound diagnostic result. In general, the lack of soundness arises. Both, obtaining an ideal diagnostic result and reducing the computational complexity, is obtained by the partially coordinated diagnostic algorithm.

Includes subconference "Prestigious Applications of Intelligent Systems (PAIS 2008)."