Lothar Litz

Formal methods in PLC programming

A detailed generic model of the control design process is introduced and discussed. It is used for surveying different formal approaches in the context of PLC programming. The survey focuses on formal methods for verification and validation (V&V). The varying works in this area are categorized using three criteria: the general approach (A) to the task (model based, constraint based or without a model), the formalism (F) (Petri net, automata, etc.,) used to state the formal description, and the method (M) (model-checking, reachability analysis, etc.,) used to analyze the properties. Based on these three criteria (A-F-M) a three letter code for V&V approaches is introduced. Some works from the multitude of V&V research are presented and categorized using this new system.

Reduction of fuzzy control rules by means of premise learning - method and case study

Rule number reduction is important for fuzzy control of complex processes with high dimensionality. It is stated in the paper that this issue can be treated effectively by means of learning premises with general structure. Since conditions of rules are generalised by a genetic algorithm (GA) rather than enumerated according to every AND-connection of input fuzzy sets, a parsimonious knowledge base with a reduced number of rules can be expected. On the other hand, to give a numerical evaluation of possible conflicts among rules, a consistency index of the rule set is established. This index is integrated into the fitness function of the GA to search for a set of optimal rule premises yielding not only good control performance but also little or no inconsistency in the fuzzy knowledge base. The advantage of the proposed method is demonstrated by the case study of development of a compact fuzzy controller to balance an inverted pendulum in the laboratory.

FAULT DETECTION OF DISCRETE EVENT SYSTEMS USING AN IDENTIFICATION APPROACH

In this paper, we focus on the identification of large-scale discrete-event systems for the purpose of fault detection. The properties of a model to be useful for fault detection are discussed. As appropriate model basis the nondeterministic autonomous automaton is chosen and metrics to evaluate the accuracy of the identified model are defined. An identification algorithm which allows setting the accuracy of the identified model is presented. Results are given for two case studies, one of a laboratory and another one of an industrial plant.

Verification and validation of control algorithms by coupling of interpreted Petri nets

Interpreted Petri nets (IPN) are an extension of the basic Petri net (PN) framework. They are ordinary PN with additional components to model the information flow to and from the PN. This contribution gives a general overview of the control design process. Furthermore, it shows how the different tasks in this process, especially analysis, verification and validation, can be successfully solved within the formal IPN framework.

Black-box identification of discrete event systems with optimal partitioning of concurrent subsystems

This paper proposes a data-driven method to determine concurrent parts in Discrete Event Systems (DES). The aim is to improve the results of black-box identification methods without considering any system information except of observed data. To allow an analysis of the collected data, the impact of concurrency on the exhibited system data is determined by two criteria. We propose to use an optimization algorithm that isolates concurrent parts of the system by minimizing concurrency expressed by the two proposed criteria within the determined subsystems. A lab-size application shows the potential of the method for real-world manufacturing systems. The aim is to deliver optimal identified models for fault detection and isolation.

Learning premises of fuzzy rules for knowledge acquisition in classification problems

Abstract. A key issue in building fuzzy classification systems is the specification of rule conditions, which determine the structure of a knowledge base. This paper presents a new approach to automatically extract classification knowledge from numerical data by means of premise learning. A genetic algorithm is employed to search for premise structure in combination with parameters of membership functions of input fuzzy sets to yield optimal conditions of classification rules. The major advantage of our work is that a parsimonious knowledge base with a low number of rules can be achieved. The practical applicability of the proposed method is examined by computer simulations on two well-known benchmark problems of Iris Data and Cancer Data classification.

NCS testbed for ambient intelligence

Ambient intelligence (AmI) aims to make the environment of human beings more intelligent and sensitive in order to enhance the quality of life. In AmI systems a large number of sensors, actuators and controllers are distributed on low power and low performance nodes communicating via wireless ad-hoc networks. These systems build a special class of networked control systems (NCS), where many constraints exist such as limited bandwidth. In order to develop control algorithms under AmI constraints, a testbed is required which takes them into account. First, this paper presents a simulation environment that offers a library of NCS blocks. Thus, the constraints can be considered and integrated in the design process. Second, it describes a real process, an inverted pendulum, which is automated based on Mica nodes. These nodes were designed especially for AmI purposes. This real NCS serves as a challenging benchmark for proving the AmI suitability of the controllers.

QoS-adaptive control in NCS with variable delays and packet losses - a heuristic approach

Networked control systems (NCS) have become a widely considered research topic. Two main challenges relate to variable network delays and to packet losses. This paper describes an approach considering this crucial quality of service (QoS) parameters at the same time. The solution is based on a heuristic approach and its effectiveness is outlined in an example. The resulting control algorithm explicitly depends on current measurements of the network QoS parameters. They are used to adapt the control parameters and the sampling rate, as well. In order to design and evaluate the control algorithm, a simulation environment based on MATLAB/Simulink is presented that is able to model the main NCS effects.

An FDI Method for Manufacturing Systems Based on an Identified Model

In this paper a generic method for fault detection and isolation (FDI) in manufacturing systems considered as discrete event systems (DES) is presented. The method uses an identified model of the controlled process to be monitored which is built on the basis of observed fault free behavior. A special term of accuracy is motivated that helps to identify an efficient model. This paper gives an overview of the method that consists of the identification and the use of the identified model for fault detection and isolation. Furthermore, the theoretical framework of the method will be explained. Experiences of an industrial application are described to show the relevance of the method for large scale manufacturing systems in operation.

Correctness analysis of Petri net based logic controllers

Petri nets are able to express the causality as well as the concurrency of a control algorithm. Signal interpreted Petri nets (SIPN) show good properties for model logic controllers. In SIPN the firing of a transition depends on input signals from the environment and the SIPN influences the environment via output signals. Since the function of a logic controller is basically determined by software, the question of software quality arises in this area. ISO/IEC 9126 defines six characteristics of software that can be used as quality criteria. In the six characteristics are set in the framework of controller design: the functionality and reliability of the software rely on the correctness of the algorithm. In this criteria for formal correctness of SIPN control algorithms are given, and it is shown how they can be evaluated using the SIPN reachability graph. Due to a close relation between sequential function chart (SFC) according to IEC 1131-3 standard [IEC 1992] and SIPN the criteria are also valid for SFC.