Jerzy Rutkowski

Genetic-Algorithm-Based Method for Optimal Analog Test Points Selection

A new approach to an optimal analog test points selection is proposed. The described method uses ambiguity set concept and evolutionary computations to determine the optimal set of analog test points. After a brief introduction to analog testing and genetic algorithms, the proposed strategy is explained. The presented evolutionary approach is illustrated by a practical example of analog circuit and by a series of hypothetical circuits. The efficiency of the technique is compared with a method based on entropy index, and the obtained results are discussed

Application of genetic programming to edge detector design

The new approach to edge detection is presented in this paper. The proposed method uses genetic programming (GP) to search for digital transfer function of image edge detector. The found function can be easily implemented to any programmable logic device (PLD) that allows to build a fast system of image processing

Analog fault AC dictionary creation - the fuzzy set approach

This paper discusses the basic concept of analog functional test approach. Recently, most on going research has been focused on distinguishing faulty or healthy circuit - from the manufacturers point of view, this is more important than locating particular faulty element. The article shows the concept of fuzzy theory approach to functional test creation. To find and locate faulty circuit (system) the sensitivity matrix can be used. It has been observed that for most practical circuits the sensitivity matrix is sparse. This observation has been utilized in the fuzzy expert system and the system modification based on the reduction of the sensitivity matrix size has been proposed. The modified system can effectively decide whether the system is faulty or healthy. The results obtained and presented here are promising and tend to prove that the proposed strategy can improve fault detection on a production line

Barriers to introduction of e-learning: A case study

In 2005 an integrated blended e-learning service based on Moodle Learning Management System (LMS), was introduced in Silesian University of Technology, Gliwice, Poland [1]. Review of LMS content shows that some courses contain insufficient number of components, and in many cases the quality of resources is poor. The goal of our research was to identify why some teachers are reluctant to utilize LMS. A survey was carried out in order to examine to what extent teachers accept and make use of various LMS elements. Teachers were also asked to indicate main barriers to LMS-based teaching and present their comments on perspectives of blended learning. Several figures, charts and comments resulting from the survey are presented in the paper. The main obstacles were identified and the suggestions how to eliminate them are presented in the paper.

Optimization of PWL analog testing excitation by means of genetic algorithm

The method presented in this paper uses evolutionary computations to analog circuit testing stimulus shape optimization. The proposed testing excitation is a PWL waveform that generates responses of the tested circuit with appropriate relationship between current values of its specifications and the value of the observed parameter. The energy of error signal between nominal and actually obtained responses is designated for the analyzed parameter. The fitness function applied to the evolutionary system allows to optimize the separation level of the identified circuit states and to improve linearity for relationship between the energy level and the circuit state.

The Influence of Global Parametric Faults on Analogue Electronic Circuits Time Domain Response Features

This paper presents an analysis of an influence of global parametric faults (GPF) on analogue integrated circuits (AIC) time domain (TD) response features, such as overshoot, delay time, rise time, maxima and minima, first differential maxima and minima. The novel approach is the analysis of relations and superrelations between features which are discussed in details. The presented results should increase testability and diagnosability of the global parametric faults on a production line of analogue and mixed electronic circuits. The observed AIC faults belong to parametric ones and are caused by technological process. GPF influence on aforementioned features is presented with the use of an exemplary circuit. A research presented in this paper may be used as an introduction to a fault driven test (FDT) diagnosis with the use of simulation before test (SBT) methods.

PCA application to frequency reduction for fault diagnosis in analog and mixed electronic circuit

This paper is focused on the diagnosis of analog and mixed electronic circuit in frequency domain. A multi tone signal excites a circuit under test in order to detect and locate a single fault. The input signal consists of a number of frequencies which are selected by Principal Component Analysis (PCA) algorithm. An exemplary circuit has been examined and results have been discussed.

Simulated Annealing with Fuzzy Fitness Function for Test Frequencies Selection

This paper discusses the optimization of input source frequencies to obtain the highest detection and location rate of a circuit under test. The minimum number of frequencies for distinguishes between fault and fault free elements are desired. We briefly describe the basic concepts of simulated annealing method and fuzzy evaluation (fitness) system. Presented application shows how to code the optimization problem. It has been compared algorithm with fuzzy goal function and weighted goal function. Obtained results are attached and discussed.

A two stage neural network DC fault dictionary

A new concept of neural network DC fault dictionary is presented. The neural network is combined with two networks: Learning Vector Quantization network and Multilayer Perceptron. Simulation results are presented and they show that the proposed approach is an improvement as compared with classical nearest neighbour rule approach.<<ETX>>

Multidimensional search space for catastrophic faults diagnosis in analog electronic circuits

This paper describes new method of fault diagnosis in analog electronic circuits (AEC). Fault diagnosis in analog electronic circuits is in general tested along with one dimension: the generator frequency. In this paper, we present a novel approach that uses more than one dimension to test AEC (those new dimensions are: the load resistance & reactance, generator resistance & reactance). Proposed novelty can be perceived as a hardware solution of tester of AEC – old approach: tester with one dimension search space, new approach: tester with multidimensional search space. A search space is defined as the set of input variables. In this article, we focus on how an increasing number of dimension of search space (in analog circuit diagnosis) can influence on the identification of states of circuit under test (CUT). The criterion of optimization is the maximum number of unequivocally identified states of CUT, and for that purpose a heuristic algorithm that uses Particle Swarm Optimization (PSO) is used. Some new terms are introduced like: group, testing strategy, testing path. The proposed method has been checked for single catastrophic faults of CUT.