Trends in Contemporary Methods for Digital Signal Processing

E. Vladkov

Key Words: Digital Signal Processing; digital signal processors; Harvard architecture; fixed point processors; floating point processors.

Abstract. The main application areas of digital signal processing methods and the software and hardware considerations are discussed. General conclusions are made for the use of assembly and high level languages in digital signal processing algorithms. Criteria for the choice of programming language are formulated. The special features and advantages of the Digital Signal Processors (DSP), compared to the PC-based solutions, are considered. Criteria for the choose of fixed or floating point format digital signal processors are suggested.




Conception for Building of Control Systems for Flexible Assembly Automation

Ò. Neshkov

Key Words: Assembly automation; flexible manufacturing systems; control systems for manufacturing; robotics for assembly.

Abstract. The paper presents the experience of the Department of Discrete Production Engineering at the Technical University of Sofia in the field of control systems for assembly automation.The approach of building up control systems for robotized assembly structures and the conception of control systems for a cell are discussed. Presented are also some of the results of the gathered experience.


Internet Technologies and Databases for Presentation of Production Companies

T. Neshkov, I. Pankov, J. Pankov

Key Words: IT; WEB design and WEB presentation; Internet for Business application.

Abstract. This paper is directed to the problems linked with the presentation of production and trade enterprises over the Internet. The whole process from collecting, classification and searching to presenting data for manufacture catalogues is observed.Six different models for unified collecting and classification of data from production companies are commented. Each model is oriented to a different production line of the firms. The models differ form a very basic, pure textual model to complex graphical presentations with a wide range of miscellaneous parameters. A specialized Database with a number of searching facilities is viewed. One critical factor for Web-oriented databases is working time of online Database quires. Increasing the amount of preliminary data processing can minimize this time. Three different approaches are shown. The first one is a pure full-text search without any introductory preparations; the next one is based on a simple preliminary word extracting as a base for subsequent online queries. The last one is more complex - extracting of words followed by their classification as terms by branches. Analysis is made and the link between speeding up online queries and increasing the time for pre processing of the data is detected. Future directions for expansion of the Internet concerning the small- and medium sized enterprises (SME) concludes this article.


Thematic Network COFURN ­ New Ideas and Solutions for Integration in Furniture Industry

I. Bachkova

Key Words: Standardization; product data model; ISO-10303; e-commerce; furniture industry.

Abstract. Main aim of this paper is to present the Thematic Network COFURN. With the purpose to disseminate the European initiative for creating Networks in the field of standardization, the main premises, goals and structure of COFURN are presented. Some knowledge information about the Application protocol - AP 236 for furniture industry sector and a platform for neutral data format are proposed.


Improving Noise Immunity of Speech Recognition Systems

R. Ivanov

Key Words: Robust speech recognition; speech enhancement; noise reduction.

Abstract. The development of robust automatic speech recognition (ASR) systems that maintain a steady and high level of recognition accuracy in dynamically varying adverse condition (channel and acoustical environment distortions) is one of most important factors that determine the success of speech recognition applications. The most of current ASR systems that are designed to be speaker independent work well only in relatively quiet office environment and they can perform very poorly when are tested in different environment. Commercial applications such as speech recognition over telephone (radio) lines and Internet (information access and extraction); hand-free GSM (mobile) phones etc. demand an even greater degree of channel and environmental robustness. The goal of this survey is to provide review of methods and techniques for robust speech recognition.


Neural Networks for Analysis of Organizational Behaviour in Industrial Systems: I. MLP with the Backpropagation Algorithm

G. Mengov, G. Sotirov

Key Words: Neural networks; multilayer perceptron; organizational behaviour; work motivation in industrial systems.

Abstract. This paper investigates the applicability of artificial neural networks to analysis of empirical data from studies of organizational behaviour in industrial systems. In part (I) the multilayer perceptron trained with the back-propagation algorithm is evaluated against linear regression models on the example of a sample of 1104 industrial employees representative for Bulgaria in 1994. It is demonstrated that neural networks account for the data non-linearity better than regression. The effects of noise and deviation from the normal distribution in the data pose natural limitations to both methods’ prediction capabilities.


Flexible Object-Oriented Upgrade of a Generalized Model Database

A. Rozeva

Key Words: Database design; object-oriented upgrade; flexibility of information structures.

Abstract. Most of databases nowadays work in environments that are quickly evolving, irregular and unpredictable. They need to be flexible enough and able to share their information content for providing database integration. The software approach that proved to be most suitable and attractive in contemporary database design is the object-oriented one. It has been accepted as a basis for developing an upgrade level over a database with conceptual model called generalized for the sake of adapting it to the needs of various application environments. Additional techniques have been integrated therein for raising the flexibility of information structures in data intensive and often changing environments.


Algebra of Algorithms in Procedural and Object-Oriented Structured Programming

G. Istatkova

Key Words: Compositional algebra of algorithms; Boolean algebra of algorithms; algebraic methods for testing; verification and structural transformation of algorithms; procedural and object-oriented structural programming.

Abstract. In this parer, algebraic method for transformation procedure-oriented compositional models of structured algorithm to functional multiple t hreads form, independent to module form in program paradigms is proposed. Method is based on analysis of functional threads in disjunctive logical model of algorithms in Boolean algebra, isomorphic to procedure-oriented compositional algebra of acad. Gluschkov. To convert algebra of acad. Gluschkov to Boolean algebra of algorithms logical forms of basic control structure are used. Static verification and testing of algorithms using disjunctive model is offered. The testing of complex hierarchical program system using functional models of algorithms is compared in procedural and object-oriented structural programming. Testing in procedural programming is considered as “P-problem”, that can’t be resolved in practice with functional system complexly increase. It is shown that class in object-oriented paradigm resolves this problem and allows efficient program interpretation of hierarchical functional libraries on classic fon Neuman computer architecture.


Control of Robot Manipulators Using Composite Reduced-Order Observers

G. Lehov

Key Words: Robot control; observer-based control; robust stability; Lyapunov’s direct method.

Abstract. A robot control scheme with a linear composite reduced-order observer is presented in this paper. The resulting closed-loop system is proven to be uniformly ultimately bounded in the trajectory tracking case, and exponentially stable in the constant set-point case. Relations for selecting the controller gains are derived on the basis of the stability analysis. Simulation results demonstrating effectiveness of the control scheme are presented.


Composition of Multivalued Logical-probability Functions and their Application in Multistage Stochastic Systems

E. Gegov

Key Words: Multivalued logic; probability and interval uncertainty; Multivalued Logical Probability Functions (MLPF); composition.

Abstract. . A class of multivalued logical-probability functions (MLPF) for modelling processes in multifactor nonlinear stochastic plants and systems is proposed. A method of composing MLPF is developed by means of multivalued logic and probability theory. The method can be used in multistage technological, power, and other types of systems. An example from the field of mineral technologies is presented.



Power Plant Advanced Control Systems Efficiency

M. Hadjiski

Key Words: Advanced control systems; decision rules; efficiency estimation.

Abstract. The possible benefit sources of implementing advanced control systems are analyzed. Decision rules for control strategy and control methods are proposed. The main point is optimization of the ratio “profit/investments” during the implementation. The problems of real e fficiency estimation for control system after the implementation are treated. Examples of power steam generators are provided.


Wireless Technology Using Code Division Multiple Access

H. Turlakov, K. Boyanov

Key Words: Wireless communication; mobile communication; multiplexing; GSM; CDMA; TDMA; FDMA.

Abstract. In survey are described the main principles of code division multiple access (CDMA). The comparison between existing wireless technologies is given. The advantages and drawbacks are also discussed.


Neural Networks for Analysis of Organizational Behaviour in Industrial Systems: II. Artmap-type Neural Networks

G. Mengov, G. Sotirov

Key Words: Neural networks; ARTMAP; organizational behaviour; work motivation in industrial systems.

Abstract. This second part of the paper introduces the ARTMAP family of neural networks as an alternative tool for analysis of mutidimensional patterns of psychological interactions in industrial organizational settings. An architecture of dARTMAP and Fuzzy ART is proposed, which augments the properties of the neural network in understanding the empirical data. The neural network is tested on a questionnaire-based sample from Bulgarian industrial organizations, whereby its qualities are evaluated and potential benefits are suggested.


Application of Microcomputer Spreadsheet Programs in Time-Domain Modeling and Analysis

E. Manolov

Key Words: System level modeling; time-domain analysis; computer simulation; computer spreadsheet programs; Excel.

Abstract. The possibilities of applications of the microcomputer spreadsheet program Excel in time-domain analysis of linear and non-linear analog and discrete systems are investigated and demonstrated. To illustrate this potentiality two examples are modeled and examined: simple linear discrete system (first order circuit) and non-linear analog system (Van der Pol equation for oscillator). The using of the microcomputer spreadsheet Excel in the shown applications gives an opportunity of making sophisticated computations and analyses by applying various numerical methods and allows direct control and visualization at every step of the problem solution.


Improving Noise Immunity of Sound Localization Systems with Microphone Array

M. Momchedjikov, A. Nikiforov

Key Words: Microphone array; sound localization systems

Abstract. The evaluation of signal-to-noise ratio of microphone array in contrast with single microphone system is offered in the paper. Made analysis brings about conclusion: the beamforming capabilities of microphone array systems allow highly directional sound capture, providing superior signal-to-noise ratio when compared to single microphone performance.


MRWV - A Design Model for Real World Visualization

E. Racheva

Key Words: Databases; image databases; real world; information systems; GIS; multimedia information systems and HPPM.

Abstract. The visualization of the Real World (RW) or Real Geographical World (RGW) is a very difficulty process. The main reasons for this problem are the three dimensionality of the RGW and the infinity of it.This paper presents a Design Model of Real World (MRWV). It is based on the Hierarchical Point to Point Model (HPPM), which was presented in Budapest European Commission 4th Geographical Information Systems (GIS) Workshop in 1998. It is a hierarchical “Point to Point” system. Each Point is an Object or Set of Objects of the RW.In the paper is showed the mathematical structure of the hierarchical system of the MRWV. The Model is really created on the Points if they are situated, for example, on the 3D RGW map. On the other hand, HPPM has created in two variants with two or three-dimensional structures. It is very useful for the visualization in the 2D GIS and Multimedia Information Systems (MMIS).


Control System of Hybrid Device for Automatic Assembly

V. Vitliemov, S. Belev, T. Hristova

Key Words: automatic assembly; passive-active method; force-speed control.

Abstract. In the recent paper is presented a control system of hybrid device for automatic assembly of peg-hole. The structure of control system is with quazzy-continuous action that decreases the duration of assembly process. The developed control system consists of inner loop in ”control by velocity” regime and outer loop in ”control by force” regime.


Image Coding in CAD/CAM System for Knitting Industry

E. Zaharieva, N. Siniagina

Key Words: CAD/CAM system; Image Coding; Knitting Industry.

Abstract. This paper treats the problems of image coding in CAD/CAM systems for knitting industry. Image coding criteria are: minimum of used memory; facility and accuracy of design activities. Knitting objects are classified in this paper, image coding methods are represented for describing image objects as CAD/CAM system data. These methods are applied in CAD/CAM system for cotton knitting machine design.