Showing papers presented at "Computer Aided Systems Theory in 1991"

Multidimensional Constant Linear Systems

Abstract: A continuous resp. discrete r-dimensional (r≥1) system is the solution space of a system of linear partial differential resp. difference equations with constant coefficients for a vector of functions or distributions in r variables resp. of r-fold indexed sequences. Although such linear systems, both multidimensional and multivariable, have been used and studied in analysis and algebra for a long time, for instance by Ehrenpreis et al. thirty years ago, these systems have only recently been recognized as objects of special significance for system theory and for technical applications. Their introduction in this context in the discrete one-dimensional (r=1) case is due to J. C. Willems. The main duality theorem of this paper establishes a categorical duality between these multidimensional systems and finitely generated modules over the polynomial algebra in r indeterminates by making use of deep results in the areas of partial differential equations, several complex variables and algebra. This duality theorem makes many notions and theorems from algebra available for system theoretic considerations. This strategy is pursued here in several directions and is similar to the use of polynomial algebra in the standard one-dimensional theory, but mathematically more difficult. The following subjects are treated: input-output structures of systems and their transfer matrix, signal flow spaces and graphs of systems and block diagrams, transfer equivalence and (minimal) realizations, controllability and observability, rank singularities and their connection with the integral respresentation theorem, invertible systems, the constructive solution of the Cauchy problem and convolutional transfer operators for discrete systems. Several constructions on the basis of the Grobner basis algorithms are executed. The connections with other approaches to multidimensional systems are established as far as possible (to the author).

Minimal Realization Procedures Based on Balancing and Related Techniques

Abstract: The main advantage of minimal realization procedures related to balancing is that the determination of the order of the minimal realization relies on a single, numerically reliable, rank decision Moreover, usually the same procedures can be also used to compute reduced order approximations of the given system The paper presents eight minimal realization methods related to balancing techniques and compares their numerical performances New derivations for some existing methods are provided and several new methods are suggested The performed numerical tests confirm the superiority of recently developed square-root methods over other existing methods

Bus Protocol Conversion: from Timing Diagrams to State Machines

Abstract: A designer of hardware systems always has to deal with the somewhat standard task of interfacing system components to a system architecture using different communication protocols. Still there exists little computer aid to support the design of integrated interface circuits today. We present a high level synthesis method to derive interface controller units from the timing diagram specifications of the participating communication protocols. Transformation from timing diagrams to terms of a process calculus provides a basis to master the conversion problem. If A and B are process behavior terms that represent timing diagrams, a third specification X can be derived automatically, where X describes a converter process for A and B. X exists, if there is a solution for the equation A·X=X·B, where “·” is the concurrency operator of the process calculus. The operational semantics of process calculi permits a state-oriented view of the derived converter specification X. From here it is a well-known procedure to synthesize a hardware realization.

CAST Tools for Automatic FMS-Dynamics Models Synthesis

Abstract: A general approach to automatic synthesis of FMS-dynamics models is presented. The problems that require analytical solution are discussed. An illustration of the approach with respect to a certain class of manufacturing processes and using Petri nets as a modelling formalism is given. Application of the synthesis method is indicated.

Linear and Algorithmic Formulation of Co-operative Computation in Neural Nets

Abstract: In this paper we consider a formulation of neural computation in terms of multilayer co-operative processes between local inference units. Each neuron is described in terms of a local computation frame with slots and semantic tables to represent the meaning of the input and output representation spaces. After some methodological remarks on the distinction between processors and processes in the auditor paradigm, we introduce the computational model of co-operativity where global computations are divided into a large number of dinamically reconfigurable sub-computations. Some theoretical requirements are considered and the lack of developments on concurrent programming is pointed out. Then, the frame neuron is introduced and the meaning of the different slots is stablished. Linear and algorithmic formulation appears as a particular case of the frame to which we can reduce all previous connectionistic formulations. This computational model is used to “revisit” the analogical level where neuronal function is formulated in terms of linear spatio-temporal filters followed by a threshold function.

Geometric Theory of Nonlinear Dynamical Networks

Abstract: In this paper we start with some historical remarks about nonlinear network theory and motivate the geometrical approach of nonlinear dynamical networks by means of a simple example. Then we outline the theoretical foundations of this theory and discuss some results. Finally, we refer to geometric approaches in physics and in other areas of engineerings and explain the common features and the differences between electrical network theory and classical mechanics.

Input-Output Model For Unconventional Samped-Data Control Systems

Abstract: A Block Multirate Input-Output model for sampled-data control systems has been defined by the author, by considering blocks of measurement and control samples. It allows the design of periodic and multirate controllers providing strong control performances. The aim of this paper is to extent these results to the case of unconventional sampling, and to analyse the use of the new controllers to counteract external disturbances.

Complete Transforms and Their Incidence in Artificial Perception Systems Theory

Abstract: The importance of this, theorem rests in that it integrates the algebra and analysis of the data fields, stating that, given an algebraic partition with certain restrictions, them the computation of a number of analytical descriptors independently in each partition provide for a complete and non-redundant description of the data fields.

Automatic Robot Programming by CAST

Abstract: A comprehensive framework for generating a robot's program for an automated production system will require an integration of several layers of system theory-based support methods and tools. Each layer of the robot's program synthesis system requires different CAST tools. The tools for each level are: Level 1 :graph search methods Level 2 :Petri net methodology Level 3 :discrete dynamical system methods Level 4 :discrete optimization methods Level 5 :event based system formalism

General Systems Theory Requirements for the Engineering of Complex Models

Abstract: We live in a period of time where caused by the availability of large scale systems for transportation and communication strong interconnection of states of our local environment exist together with states of the global environments. Locally working systems have usually been grown in a natural evolutionary way. This applies not only to technical systems but also to socialor economic systems, which exist today. The liberation of trade, travel, cultural exchange and scientific results between the different countries is the main reason for such a strong interconnection of local states and for the existence of large scale systems.

User Interface to a CIM-Database

Abstract: “Low Cost” modular CIM concepts especially for small and medium sized companies will be of great interest in future In this paper a new module of the user interface between the users at the CAD working place and the CIM database and the main tasks of the interface are described The “low cost” modular CIM solution is established in order to embed all its components into the existing infrastucture of the company, according to their demands Therefore physical and logistic connections of the existing PPS system as well as the installed CAD workstations with the working places have to be established These parts have to be included into the superior systems of recording data of the plant and of operating The realization of this solution is accomplished by using the C — programming language, Informix database and Informix — SQL as database query language, in order to create the custom applications with database management With these tools, the structure of a relational database and the ANSI — standard Structured Query Language (SQL) are provided

A Transformation of the Problems of Minimal Satisfaction of Constraints

Abstract: This paper concerns the search problems, called the problems of minimal satisfaction of constraints. The problem transformation method is proposed. It is proved that if a problem P2 is a homomorphic image of a problem P1 then P1 is transformable into P2. Next, a relationship between a computational complexity of two problems P1 and P2, being in the problem transformation relation, is investigated. A main result of this paper asserts that if P1 is polynomially and homomorphically transformable into P2 and P1 is NP-difficult then P2 is also NP-difficult.

An Integrated Environment for Computer-Aided Control Engineering

Abstract: This thesis considers the construction of a system to support the total design cycle for control systems. This encompasses modelling of the plant to be controlled, specification of the final objectives or performance, design of the required controllers and their implementation in hardware and software. The main contributions of this thesis are : its development of a model for CAD support for controller design, evaluation of the software engineering aspects of CAD development, the development of an architecture to support a control system design through its full design cycle and the implementation of this architecture in a prototype package. The research undertakes a review of general design theory to develop a model for the computeraided controller design process. Current state-of-the-art packages are evaluated against this model, highlighting their shortcomings. Current research to overcome these shortcomings is then reviewed. The software engineering aspects to the design of a CAD package are developed. The characteristics of CAD software are defined. An evaluation of Fortran, Pascal, C, C++, Ada , Lisp and Prologue as suitable languages to implement a CAD package is made. Based on this, Ada was selected as the most suitable, mainly because of its encapsulation of many of the modern software engineering concepts. The architecture for a computer-aided control engineering (CACE) package is designed using an object-oriented design method. This architecture defines the requirements for a complete CACE package including control-oriented data structures and schematic capture of plant models. The details of a prototype package using Ada are given to provide detailed knowledge in the problems of implementing this architecture. Examples using this prototype package are given to demonstrate the potential of a complete implementation of the architecture.

Decision Supported Systems and Their Industrial Applications

Abstract: The adjustment of a hollow section rolling line in the VOEST-Alpine Krems can be made more efficiently by use of a decision supported system Basic requirement is a complete knowledge base of the whole process There exist different methods how to get this knowledge base and how to treat it Using a learning component, the system will be able to enlarge it's knowledge by itself New logical languages like FORLOG make the programming easier, especially a bidirectional data transfer between logical and procedural languages is possible

Systems Optimization on Retinal Research

Abstract: The analisys of the properties of specialized cells in the retina using concepts from Systems Theory allow us to exclude several retinal mechanisms in the generation of responses to spatial and temporal stimuli, suggesting other mechanisms as more probable We need more experimentation to clarify the underlying mechanisms

A High Level Interactive Design Environment for Complex Systems

Abstract: CASSY is a system for formal representation and verification of architecture proposals at the system level. The designer describes the analog and digital signal in either the time or frequency domain using the high-level representation language SIBA. The system components are specified hierarchically with the high-level language KOBA. Then HISIM performs a symbolic and formal simulation of the specified system, avoiding point to point processing of e.g. sample streams.

The Signal Description Language SIBA

Abstract: SIBA is inspired by the description style of an experienced designer. Signals for instance are not represented by a list of amplitude values over time or frequency, but as a symbolic description of their waveforms. SIBA describes both digital and analog signals. As usual in communication engineering, signals can be described in frequency domain as well as in time domain. In SIBA signals can be combined and manipulated by arithmetic operations or control structures to form new signals in nested description blocks. The user can define its own signals to form a hierarchical description. The application of SIBA motivated the development of special tools like ONSA, AUCA or the help system. Moreover the communication between the user and the language is supported by an ergonomics graphical user interface named CASSY-X. SIBA, the related tools and the user interface constitute a compound which is a powerful system for the description an generation of signals.

Incorporating Heuristic Rules in a Multifaceted System

Abstract: Rule-based systems have become very popular because of their interesting features like explanation facilities, possibility of capturing imprecise knowledge, a simple implementation style, etc. A major disadvantage, however, is their impotence in handling complex systems.

CAST.FSM Applied to VLSI Synthesis: Experimental Results and Requirements for Industrial Use

Abstract: Castfsm is a method bank for the treatment of finite state machines In this paper we report on the practical application of Castfsm to the controller synthesis as part of the VLSI synthesis We present experimental results for the algebraic decomposition and the shift register realization of the MCNC benchmark FSMs and controllers of digital circuits Requirements for further developments with respect to industrial applications are discussed for the two specific methods used in this work as well as for CAST method banks in general

Systems Modeling and Process Control

Abstract: With the rapidly growing complexity of software systems, systems engineers and users are often confronted with a constantly increasing variety of software tools which have to be integrated and operated. Since most systems offer no or little support for these tasks they have to be handled by well-trained specialists. This is a serious bottleneck limiting the advantages of those systems.

Experience in the Development of Advanced Modelling Environments

Abstract: The basic features and a possible software architecture of an integrated simulation environment are presented in the paper. Such an environment should support the whole modelling process starting with the identification of model structure and finishing with the evaluation of alternative model formulations. Based on the experience of five working prototypes in the time span of four years, a general framework is described which contains both a quantitative and a qualitative simulation approach.

Object Identification and Retrieval in a CAST Library

Abstract: The classification and recognition problems are fundamental to an understanding of object identification and retrieval in a CAST library or repository.

CA-Methods for the Development and Construction of Hydroelectric Power Plants

Abstract: This paper reviews the activities of the Institute for Waterpower and Pumps in the use of CA-methods for the development and construction of Hydroelectric Power Plants. After discussing the role of CAST/CASE in the problem solving process and explaining the methodologies in modelling and designing of the power plant, we show typical examples of solutions out of the software packages, which have been developed at the Institute.