Intelligent Systems Techniques Research Articles

A robust on‐line learning algorithm for intelligent control systems

AbstractThis paper describes a novel error extraction approach for exploiting the strength of Levenberg–Marquardt (LM) optimization technique in intelligent control systems. Since the target value of the control signal is unknown, tuning of the controller parameters becomes a tedious task if the knowledge about the system and the environment is limited. The suggested methodology utilizes the sliding model control (SMC) technique. The error extraction scheme postulates the form of error on the applied control signal using the discrepancy from the prescribed reaching dynamics. The devised approach has been tested on the non‐linear Duffing oscillator, which has been forced to follow a periodic orbit radically different from the natural one. The results obtained through a series of simulations have confirmed the high precision and robustness advantages without knowing the analytical details of the system under investigation. The issues of observation noise and the stability in the parametric space have approximately been addressed from the point of SMC perspective. Copyright © 2003 John Wiley & Sons, Ltd.

Soft computing framework for intelligent human-machine system design, simulation and optimization

This paper proposes a novel soft-computing framework for human–machine system design and simulation based on the hybrid intelligent system techniques. The complex human–machine system is described by human and machine parameters within a comprehensive model. Based on this model, procedures and algorithms for human–machine system design, economical/ergonomic evaluation, and optimization are discussed in an integrated CAD and soft-computing framework. With a combination of individual neural and fuzzy techniques, the neuro-fuzzy hybrid soft-computing scheme implements a fuzzy if-then rules block for human–machine system design, evaluation and optimization by a trainable neural fuzzy network architecture. For training and test purposes, assembly tasks are simulated and carried out on a self-built multi-adjustable laboratory workstation with a flexible motion measurement and analysis system. The trained neural fuzzy network system is able to predict the operator's postures and joint angles of motion associated with a range of workstation configurations. It can also be used for design/layout and adjustment of human assembly workstations. The developed system provides a unified, intelligent computational framework for human–machine system design and simulation. Case studies for workstation system design and simulation are provided to illustrate and validate the developed system.

Data mining techniques for improved WSR-88D rainfall estimation

The main objective of this paper is to utilize data mining and an intelligent system, Artificial Neural Networks (ANNs), to facilitate rainfall estimation. Ground truth rainfall data are necessary to apply intelligent systems techniques. A unique source of such data is the Oklahoma Mesonet. Recently, with the advent of a national network of advanced radars (i.e. WSR-88D), massive archived data sets have been created generating terabytes of data. Data mining can draw attention to meaningful structures in the archives of such radar data, particularly if guided by knowledge of how the atmosphere operates in rain producing systems. The WSR-88D records digital database contains three native variables: velocity, reflectivity, and spectrum width. However, current rainfall detection algorithms make use of only the reflectivity variable, leaving the other two to be exploited. The primary focus of the research is to capitalize on these additional radar variables at multiple elevation angles and multiple bins in the horizontal for precipitation prediction. Linear regression models and feedforward ANNs are used for precipitation prediction. Rainfall totals from the Oklahoma Mesonet are utilized for the training and verification data. Results for the linear modeling suggest that, taken separately, reflectivity and spectrum width models are highly significant. However, when the two are combined in one linear model, they are not significantly more accurate than reflectivity alone. All linear models are prone to underprediction when heavy rainfall occurred. The ANN results of reflectivity and spectrum width inputs show that a 250-5-1 architecture is least prone to underprediction of heavy rainfall amounts. When a three-part ANN was applied to reflectivity based on light, moderate to heavy rainfall, in addition to spectrum width, it estimated rainfall amounts most accurately of all methods examined.

A Multiagent System for Monitoring Industrial Gas Turbine Startup Sequences

This paper introduces a novel technique for condition monitoring of gas turbine start-up sequences. The vast amount of data and the complex processes behind on-line fault detection indicate the need for an automated solution. A multi-agent system that views the problem as the interaction of simple independent software entities for effective use of the available data is presented. The overall solution is derived from the combination of partial solutions provided by the components of the multiagent system. As a consequence, data interpretation is achieved by converting the data into appropriate information and combining individual agents' information, resulting in an automatic fault diagnosis for the engineers. This multiagent system can employ various intelligent system techniques and has been implemented using the ZEUS agent building toolkit.

Fault location in electrical power systems using intelligent systems techniques

In this work an artificial neural network based methodology is proposed for power systems fault location. Several artificial neural networks are employed, each of them being responsible for detecting faults involving a limited number of components. The proposed methodology is tested using a test system and a real Brazilian system. Indexes based on the Hamming distance are also proposed for feature selection and analysis.

Time-varying contingency screening for dynamic security assessment using intelligent-systems techniques

A time-frequency-based approach for contingency severity ranking and rapid stability assessment is described. The aim is to classify correctly all single or multiple contingencies that may result in loss of voltage or frequency stability in the first 20 s following the last disturbing action. We start by selecting a number of strategic monitoring buses where the phasor measurement units are located to capture representative voltage magnitudes and angles during detailed time-domain simulations, which cover special protection systems and on-load tap-changers. The short-time Fourier transform is then dynamically applied to the responses for extracting selected decision features as the simulation time evolves. It is shown that frequency-domain features such as the peak spectral density of the angle, frequency and their dot product evaluated over the grid areas are reliable time-varying stability indicators that can form the basis of an entirely secure classification system able to respond within 2 to 3 s after the last event in the contingency. This allows early termination of about 60% of permanently stable simulations. Fuzzy logic and neural networks are used together to make initial decisions which are then mixed by voting in order to improve the assessment reliability and security at the expense of a reduced yield. The proposed DSA scheme is successfully tested with 1027 contingencies from two widely differing test systems: a 67-bus fictitious system and a 783-bus system in actual use at Hydro-Quebec's operations planning department.

Intelligent control of a cryogenic cooling plant based on blackboard system architecture

Intelligent system techniques have been rapidly assimilating into process control engineering, with many applications reported in the last decade. Intelligent control is bringing a new perspective as well as new challenges to process control. In this paper, a software architecture for a Blackboard for Integrated Intelligent Control Systems (BIICS) is described. The system is designed to simultaneously support multiple heterogeneous intelligent methodologies, such as neural networks, expert systems, fuzzy logic, neural networks and genetic algorithms. It will be shown how such methodologies can be readily assimilated into the software architecture. The BIICS system represents a multi-purpose platform for design and simulation of intelligent control paradigms for different kinds of processes. Currently the system utilizes intelligent control techniques (neuro-fuzzy and genetic optimization) for controlling a cryogenic plant used for superconductor testing at temperatures below 100 K.

Homeostasis Compensation Processes Optimization for Intelligent Biological and Technical Systems

This paper presents the approaches enriching the diversity of homeostasis compensation processes models for natural technologies systems of the organism (NTSO). These methods use the variety of non-coalition, coalition, and cooperative games in form of stable and efficient compromises. Besides, the concepts and models obtained are used for formation of subsystem of the Intelligent technical system (ITS) extreme objective function.

Intelligent distributed simulation and control of power plants

This paper presents summaries of five research and development activities in intelligent distributed simulation and control of power plants which were presented in a panel session of the same name at the IEEE Power Engineering Society Winter Meeting on February 6, 1997 in New York City. Each of the panelists discussed methods of how they have incorporated intelligent systems techniques into their research and development efforts in power plant control. The panel was organized by the Working group on intelligent methods in station control, station control subcommittee, and the energy development and power generation committee of the IEEE power engineering society.

Utilising the student model in distance learning

A key to effective automated learning systems is the ability to dynamically adapt learning strategies to the needs of individual students. This presentation will describe the dynamic presentation facility of the DANDIE distance-based learning system that utilises a student model and intelligent system techniques to optimise presentation content and form to meet individual needs. Given a knowledge-base of course material, the system adapts the information through the selection of predefined learning strategies, dynamically constructs an interconnected set of presentation modules, and evaluates the effectiveness of the presentation strategy with respect to the individual student.

Will domain-specific code synthesis become a silver bullet?

In the commercial world, the biggest hurdles to delivering embedded intelligence in software are the software-engineering costs. We members of the intelligent-systems community have brilliant techniques available, and no end of applications is in sight, but the gap between intelligent-systems techniques and delivered software is too great. Standard programming languages and tools are too general-purpose, and the standard means of delivering a technology-as code libraries or new-fangled componentware-is simply too restricting for some specialties. One way would be through domain-specific code synthesis. Support for the development of domain-specific extensions to programming languages and of adaptive software-where code synthesis is in the inner-loop of a self-modifying programming system-would have a profound effect in this regard. This is the future of intelligent systems development.

An intelligent multi-blackboard CAD system

In this paper, automatic modeling techniques in intelligent CAD systems are discussed and a multi-blackboard based modeling method is presented. It supports modeling with multilevel accuracy and can convert abstract symbolbased description of a design object into the desired 3D model. These characteristics meet the modeling requirements for supporting the overall design process and are suitable for intelligent CAD systems. We implement them in an intelligent chair design system, AUTOCHAIR.

Fault Analysis Using Intelligent Systems

This paper examines how applications of expert systems, neural networks, fuzzy logic, and genetic algorithms are helping utilities reach their fault analysis goals and objectives. Four IEEE Power Engineering Society (PES) subcommittees or working groups are presently concerned with intelligent system applications and are coordinating their activities. A joint panel session on fault analysis was held during the 1996 IEEE PES Winter Meeting in Baltimore. The panel session dealt with various aspects of the specification, design, development, and application of different solutions for automated fault analysis using advanced intelligent system techniques and tools. In order to provide an overview of a variety of issues associated with the fault analysis automation, the panel was formed of experts from the utilities, vendors, and academia.

Intelligent aspects of biological systems

Some theoretical approaches to understand and represent intelligent updating of the DNA program and automatic knowledge processing at cellular and molecular biological level are reviewed. Limitations of the formal linguistic and computational methods are discussed vis-à-vis the advanced techniques involved in the intelligent cell system. A framework for conceptual investigations towards implementing these techniques in intelligent computer systems is presented.

Intelligent Condition Monitoring of Gas Compressor Engines

The intelligent condition monitoring of industrial plant is a rapidly expanding area of research. It combines two radically different disciplines: classical condition monitoring and intelligent knowledge-based systems. The paper presents an overview of the areas of orthodox condition monitoring systems, intelligent knowledge-based systems and intelligent condition monitoring. The paper also includes two examples of the diverse work underway in this area. Finally, there is a discussion of a current British Gas intelligent condition monitoring project, which is investigating the application of intelligent systems techniques to condition monitoring of gas compressor engines. This features a detailed summary of a particular element of the problem area. In addition the nature of the future development of this system is also discussed.

Using planning techniques in intelligent tutoring systems

This paper proposes an architecture for building better Computer-Assisted Instruction (CAI) programs by applying and extending Artificial Intelligence (AI) techniques which were developed for planning and controlling the actions of robots. A detailed example shows how programs built according to this architecture are able to plan global teaching strategies using local information. Since the student's behavior can never be accurately predicted, the pre-planned teaching strategies may be foiled by sudden surprises and obstacles. In such cases, the planning component of the program is dynamically reinvoked to revise the unsuccessful strategy, often by recognizing student misconceptions and planning a means to correct them. This plan-based teaching strategy scheme makes use of global course knowledge in a flexible way that avoids the rigidity of earlier CAI systems. It also allows larger courses to be built than has been possible in most AI-based “intelligent tutoring systems” (ITSs), which seldom address the problem of global teaching strategies.