2023 INTERNATIONAL CONFERENCE AUTOMATICS, ROBOTICS & ARTIFICIAL INTELLIGENCE (ICARAI)

3960:
Local Perturbation Limits of the Discrete Matrix Inequality Linear Quadratic Regulator Task for Differential-algebraic Systems
Andrey Yonchev
Abstract— The theory of optimal control is connected with keeping a dynamical system at minimum cost. The situation where the system behavior is depicted by a set of linear differential equations and the cost is described by a quadratic function is considered to be the linear quadratic (LQ) task. The discussed regulation task can be clearly accomplished by the solutions Q, Y of a system of linear matrix inequalities (LMIs). This paper is devoted to conditioning of the discrete-time LMI based linear quadratic regulator task for differential-algebraic (DAE) systems. To determine the local perturbation limits of the matrix inequalities we introduce an appropriate term on the right, which is somehow perturbed. Based on the perturbation analysis we obtain narrow local perturbation limits for the LMIs' solutions to the LQR task. The results are illustrated by an example.
Keywords — discrete differential-algebraic systems, local perturbation limits, LQR task, condition numbers, linear matrix inequalities synthesis
3970:
An Algorithm to Predict Hepatitis Diagnosis
Ivan Ivanov, Borislava Toleva
Abstract—We propose an effective approach and the corresponding algorithm for modeling the dataset with multiclass label of observations. We apply our algorithm to solve a classification task using random forest, decision tree, and support vector machines models. We obtain higher values of evaluated measures compared to other authors on the Egyptian Hepatitis dataset and Contraceptive dataset.
Keywords—Data Analytics, Classification Task, Resample procedure, Confusion matrix, Python.
3972:
Skin Diseases Diagnosis by Use of Optical Biopsy and Neural Networks in Case of Insufficient or Low-Quality Data
Asparuh Markovski, Tsanislava Genova, Victoria Mircheva
Abstract — In this paper, the design, training and validation of neural network for classification of skin cancers under insufficient or poor quality data using different suggested input information parameters and optimizing the number of hidden neurons is presented. An optical biopsy device developed with the participation of the authors, designed to aid the early diagnosis of skin diseases, is used. The diagnosis is based on the different response of diseased and healthy cells to illumination in specific spectral ranges.
Keywords — optical biopsy, diagnosis of skin diseases, neural networks

3973:
Explainable Artificial Intelligence in Healthcare Applications: A Systematic Review
Beatriz Costa, Petia Georgieva
Abstract—Current artificial intelligence (AI) advances and progress in medicine created a new challenge for medical AI. The” black-box” nature of AI methods has created a discussion on the use of explainability techniques to build trust and provide transparency in the AI decision-making process. A study of current state-of-the-art approaches in Explainable Artificial Intelligence (XAI) was conducted using Preferred Reporting Items on Systematic Reviews and Meta-analysis (PRISMA) research technology. In this systematic review, we provide an overview of current XAI techniques based on different taxonomies. Finally, we discuss the applications and challenges that come with the application of explainability methods in the healthcare industry.
Keywords—explainable artificial intelligence, healthcare, medical applications, machine learning
3974:
Load-bound Fuzzy Logic Control of an Industrial Nonlinear Plant
Snejana Yordanova, Milen Slavov, Desislava Stoitseva-Delicheva
Abstract—Most industrial plants have load dependent nonlinear characteristics. In the present research a Sugeno fuzzy logic controller (FLC) on the principle of parallel distributed compensation (PDC) is developed that considers the variance in the load and the plant nonlinearity. It consists of separate controllers for each boundary load and a PDC-based soft blender for their common control action as a function of the load. The controllers are designed for the boundary loads. First, PI PDC are developed to better compensate the plant nonlinearity based on, derived and validated from experimental data, Takagi-Sugeno-Kang (TSK) plant models. The PI PDC replicates the TSK plant model structure but with local PI controllers tuned with respect to the corresponding TSK local linear plant models. Second, PI controllers are designed based on the mean values of the TSK local plant models parameters. The suggested PDC load-bound controllers PI PDC and PI are applied for the control of level in a carbonisation column for soda ash production. The simulation investigations show that the PDC load-based PI PDC compensates best the system nonlinearity and improves its robustness to changes in the load compared to the PDC load-bound PI and the ordinary PDC and PI control systems designed for one load.
Keywords—Compensation of load variations, industrial carbonisation column, level control, nonlinear plant, parallel distributed compensation, TSK plant model
3975:
Binary Regresion Model for Automated Wildfire Early Prediction and Prevention
Hristina Nikova, Radoslav Deliyski
Abstract—This paper presents a model for wildfire early prediction and prevention. The model is derived by binary logistic regression using temperature, humidity, solar radiation, rain and fuel moisture conditions as initial input parameters. The input data is registered on 24/7 bases and includes 20 of the biggest fires in a 10- year period from 2007 to 2017 in a specific region. The variables in the model are presented as well as the classification table. The most significance input parameters are the fuel moisture conditions, rain precipitation as well as the maximum temperature, measured per day. The model shows very high result for true negatives and lower results for its sensitivity. The false positives and negatives are also determined. The total success rate of the model is calculated to be 84,4%. The model is tested with 3 real fires out of 15 events. The results of probability of observing fire in these 3 cases are P=78%, P=90% and P=76%.
Keywords—wildfire, data processing, binary logistic regression, prediction, prevention
3976:
Using Machine Learning Techniques to Detect Cyberattacks in Smart Homes: A Survey
Ali Sabra, Nehmeh Rmeiti, Mirna Atieh
Abstract— Smart Home solutions and the usage of Internet of Things (IoT) in home automation has become more and more popular in the past years, and these solutions become more and more complicated and the connected devices is no longer limited to some entertainment tools that improve human life, but rather now it includes some sensitive medical tools that contribute in preserving and saving people's lives, especially with regard to patients, the elderly and children which leads automatically to the rise in the number and types of connected smart devices and therefore it leads to an increase in the amount of transmitted data over the network, this issue raises the importance of the cybersecurity concept in smart homes and then increased the researchers' interest in the work to detect any anomaly traffic among smart homes networks. At the same time the smart home solutions developed and become more critical the attacks scenarios and their complexity increased too, the point that imposes on researches to use different techniques in intrusion detection (ID) and attacks classifications and the Machine Learning (ML) and Deep Learning (DL) algorithms have shown promising results in this task. This paper discusses a variety of machine learning algorithms, datasets, data collection methodologies and features selection methods that have been used to secure smart home network solutions against zero-day attacks.
Keywords— Machine Learning, Deep Learning, Cybersecurity, Cyberattack, Anomaly Detection, Smart Homes
3977:
Numerical Study on the Electrical Circuit, presented by Duffing’s Equation
Galina Cherneva, Simona Filipova-Petrakieva
Abstract—The Duffing’s equation is a non-linear second-order ordinary differential equation, describing processes with different physical nature. It analyses the status of deterministic dynamical systems. The Duffing’s oscillators can be realized by serial connection of resistor, coil and capacitor, where only the latter one is non-linear. Other two passive elements are linear. In the present paper the simplest Duffing’s oscillator is considered. On this base the dynamics of the studied electrical circuit are discussed. The Duffing’s equation solution is simulated by Matlab/Simulink model. The paper finishes with conclusion remarks about the adequacy of the proposed model to solve the Duffing’s equation.
Keywords— chaos behavior, Duffing’s equation, Duffing’s oscillator by non-linear electrical circuit
3978:
Artificial Intelligence in Creation of Scientific Written Works: Weighing the Benefits and Ethical Dilemmas - Should We Use It?
Boris Kirov
Abstract — The implementation of Artificial Intelligence (AI) within the realm of scientific writing, particularly in the natural sciences, has captured considerable interest recently. This technological advancement appears to be an instrument poised to substantially transform the production of research findings in terms of both quantity and quality. However, the stance towards the adoption of this tool remains ambivalent. While AI holds the potential to reshape the essence of writing by enabling authors to swiftly produce superior textual content, numerous concerns persist concerning various practical issues. In this paper, we examine the diverse advantages offered by the integration of AI in scientific writing and explain how it can be utilized to enhance the accuracy, consistency, and efficiency of written work in the field of natural sciences. Ethical considerations are also explored in the context of AI application in scientific writing, addressing matters related to authorship, bias, and accountability. Ultimately, we aim to summarize the benefits and drawbacks of this approach for scientific composition, determine whether any genuine ethical concerns arise from its adoption, and, of course, provide a response to the prevailing question – that is, should we employ it.
Keywords — artificial intelligence, application of AI, scientific writing, plagiarism
3979:
Classical Approach to Transient Analysis in Magnetically Coupled Electircal Circuits
Zhivko Georgiev, Atanas Chervenkov, Ivan Trushev
Abstract—The article analyzes transients in magnetically coupled circuits. The necessity to use the so-called incorrect initial conditions is justified. It is shown how to determine the initial conditions in the case when the coefficient of magnetic coupling between the inductive coupling coils is equal to 1. It is shown how an analysis of magnetically coupled coils can be carried out by studying a converted circuit without inductive couplings. Two examples with different values of the magnetic coupling coefficient M are considered. The results of the analytical solution and the numerical solution are identical.
Keywords— transients, magnetically coupled circuits, switching laws, initial conditions, “T”- equivalent circuit, transformer
3981:
Equivalent Transformations of a Cube-like Passive Linear Оne-port Networks with Equal Impedances in their Branches
Simona Filipova-Petrakieva
Abstract—In the present paper equivalent transformations of cube-like passive linear one-port networks supplied by alternating-current (AC) sources are suggested. The initial analyzed one-port is represented in the 3D space as a cube, resp. a parallelepiped. In each arm of the cube equal passive elements are connected in series so that the resulting impedance contains both real and imaginary parts, named active and reactive parts. Since all impedances are either inductive or capacitive it is not possible the phenomenon resonance to take place.
Based on the well-known transformations of passive linear one-ports networks of series and parallel type, a representation of the cube-like one-port with the only one impedance connected between the supplied nodes is proposed. One possible case of DC power supply and three possible cases of AC power supply between 3 possible pairs of nodes of the analyzed passive linear one-port network are considered.
Keywords—AC Sweep, equivalent transformations - passive one-ports, connected in series/parallel; Delta-Star transformations
3983:
Comparision between the Performance of Pneumatical and Electrical Grippers for Industrial Robots
Boris Kostov, Vladimir Hristov
Abstract— The implementation of industrial robots in the industry is mainly known with their composition in cyclic mode for positioning work. The cyclical nature of the work of the positional mechanisms allows their productivity to be defined as the reciprocal of the average cycle length. In this relationship the productivity of industrial works can be expressed as sum of two components (move time plus pause time). Time of movement is determined both by the technological process and by the choice of speed diagram of movement. The pause time can be represented as the sum of two components: The first component is related to the duration of the preparatory-final operations related to the operation of the mechanism. Such are, for example, the time for saving and releasing the load being carried, the time to enter a command about the upcoming move, etc. The second component is related to the duration of pauses in the operation of the mechanism imposed by a mandatory lock with the state or operation of other mechanisms. The present work aims to investigate the performance of a class of industrial robot using pneumatic and electrical end effector to evaluate the performance. The expectations are related to the study of the pause component in both types of grippers.
Keywords— industrial robots, speed, time, measurements, electrical gripper, pneumatic gripper
3984:
Improved Control for Fractional-order and Non-minimum Phase Integrating Plants Using a Fractional IMC and SP Structure
Anjana Ranjan, Utkal Mehta, Vladislav Slavov
Abstract— Enhanced control of integrating processes is provided in this research. The relevance of the technique may be seen in its capacity to regulate a double integrating process with time delay and non-minimal phase while dealing with large parametric uncertainties and load disturbances. A novel architecture based on the Smith predictor (SP) technique is proposed, which employs a fractional-order internal model controller (FOIMC). The required fractional-order tuning parameter is followed by the requisite gain and phase margins. The intended performance limitation is sometimes utilised to calculate another tuning parameter, such as the time constant of the fractional filter. To highlight the feature of the hybrid structure, numerical research and comparison are performed.
Keywords— Process control, Integrating plant, Fractional-order, IMC, Smith predictor, Load disturbance
3988:
Laplace Transform Method to Transient Analysis in Magnetically Coupled Electircal Circuits
Zhivko Georgiev, Ivan Trushev, Atanas Chervenkov
Abstract— In this work, transient processes in magnetically coupled circuits are analyzed using the Laplace transform method (operator method). The advantages of the operator method based on the extended Laplace transform analysis are emphasized. In this case it is only necessary to introduce the just prior to switching initial values of the currents through the coils (inductors) and the voltages on the capacitors. It is shown how an analysis of magnetically coupled coils can be carried out by studying a converted circuit without inductive couplings Two examples with different values of the magnetic coupling coefficient M are considered. The results of the analytical solution by operator method and the numerical solution are identical.
Keywords— transients, magnetically coupled circuits, switching laws, Laplace transform method, initial conditions, transformer
3991:
Robust Control Of A Multi-Operational Robotic System
Nina G. Nikolova, Georgi Stefanov
Abstract - This work proposes results of the robust control capabilities of a robust control of a multi-operational robotic system. Classical and robust systems for controlling the position of the robot-manipulator and of the current-decoupled induction motor of the technological modules have been synthesized. The systems are modelled. Their quality has been researched and comparatively analyzed.
Keywords - multi-operational robotic system, internal model robust control, robust performance

3992:
LQ/H∞ Filter Control of Axial-Piston Pump
Tsonyo Slavov, Alexandar Mitov, Jordan Kralev, Ilcho Angelov
Abstract— A modern control method for a known type of variable displacement axial-piston pumps intended for open circuit hydraulic drive systems is presented. In the developed control system the conventional hydro-mechanical controller is replaced by a digitally controlled electro-hydraulic proportional spool valve. This valve is the actuator of a swash plate which regulates the displacement volume of the pump. The generation of the control signal to the valve from a classic electronic amplifier is replaced by a real-time control system that makes it possible rapid prototyping of different control algorithms. The developed system was implemented as a laboratory test rig, on which various control laws were studied in previous works. In this paper, the results of the design, implementation, and experimental investigation of a linear-quadratic regulator with an H-infinity filter (LQ/H∞) are shown. The controller is designed on the basis of a state space model obtained through the means and methods of system identification using experimental data. The control performance of the LQ/H∞ controller was investigated not only in simulation conditions but also when controlling the pump on the laboratory bench.
Keywords— Linear-quadratic regulator, H∞ filter, real-time, axial-piston pump
3993:
Preparation and Characterization of Electrospun Nutraceptic Nanofibers from Polyvinyl Alcohol/Beeswax
Margarita P. Neznakomova, Margarita P. Neznakomova, Margarita P. Neznakomova
Abstract - In the article, it is proposed that the electrospinning process be performed with three different polymer solutions: a 9% polyvinyl alcohol (PVA) solution, PVA with an additive of natural beeswax, and PVA with an additive of natural and polyethylene wax. The electrospinning process is realized with an electrospinning chamber developed at the Technical University of Sofia. For the three spinning solutions, the change in dry matter, viscosity, and electrical conductivity was evaluated, which are directly related to the electrospinning process. The average diameters of the produced bicomponent fibers were in the range of 320–420 nm. The presented SEM images of the obtained nanomats confirm that the wax additives have an influence on the electrospinning process and, accordingly, on the diameter of the nanofibers. The area and thickness of the nanomats obtained with the addition of beeswax were evaluated. The addition of natural beeswax has been shown to increase the area of the nanomat by 16%.
Keywords— nanofibers, bioactive dressings, poly(vinyl alcohol) PVA, beeswax, electrospinning

3995:
Implementation of Digital Shadow Concept in the Field of Industrial Automation
Radovan Peter, Martin Juhás, Bohuslava Juhásová
Abstract— We are currently witnessing a gradual increase in interest in digital shadow and digital twin solutions in the field of industrial automation. These solutions require advanced approaches in the process of their creation and are gradually becoming an integral part of industrial automation. This paper focuses on the creation of a digital shadow of a physical robotic cell and progressively goes through all the phases from the analysis of communication problems through the design of possible solutions to the actual implementation of one of the proposed solutions. The connection between the virtual robotic cell and the industrial camera is based on the processing of the data sent by the camera located above the physical robotic cell. This data is processed by a program in the PLC and then transferred to the virtual environment of the digital shadow by communication between the two OPC servers. This paper provides a more detailed overview of one possible technique for connecting the virtual environment with the physical robotic cell.
Keywords—cognex, plc, abb, robotstudio, in-sight, opc, profinet, simulation, robotic cell
3996:
PID Control of the Inverted Pendulum
Kamen Perev
Abstract This paper considers the problem of inverted pendulum stabilization and performance improvement by using PID regulation. The dynamical system model is derived in the framework of the Lagrange – Euler equations and the model parameters are evaluated by using specific identification procedures. The parameter tuning of the PID controller is based on the method of dominant pole selection. Different simulations are performed, showing the influence of the third pole placement on the closed-loop system performance.
Keywords—inverted pendulum, cart – pendulum system, system parameters identification, PID control, dominant pole selection
3998:
Vehicle Assistance System Using Convolutional Neural Networks
Filip Žemla, Ján Cigánek
Abstract— The aim of this paper is to create a safety assistance system for a vehicle that is able to recognize dangerous situations while driving using the front camera. In addition to the detection of these events, the system is able to warn the driver using sound and visual effects to prevent a possible accident. The main principle of the work consists in creating and training a convolutional neural network that will be able to detect a set group of objects in images, videos or camera images captured in real time and to visualize these objects in the created application.
Keywords— neural network, security, vehicle, detection, assistance system, application
3999:
Algorithm Development for Pneumatic Actuator Control and Diagnostics
Antoniy Petrov, Albena Taneva
Abstract—This paper presents the pneumatic control system widely used in manufacturing process. This is a technology that uses the power of compressed air to create mechanical motion. The purpose of this work is to improve existing manufacturing system with simple control of pneumatic actuators. To do this is necessary to extend the simple control with adding functions to collect data for diagnostics. Such data can be used for estimating remaining lifetime of a pneumatic cylinder during preventive maintenance or replacement. Challenge in front the solution is to be compatible with several of automation systems using PLCs and easy for multiplication in laboratory equipment and real manufacturing environment and assembly systems.
Keywords—assembly system, pneumatic actuator, diagnostics, algorithm, FAS-200, S7-1200, TIA Portal
4002:
Neurophysiological Test with Use of EEG, GSR and Pulse Measurements Data for Focus Group Beverage Preference Aggregation
Stanimir Andonov, Stiliyan Georgiev, Georgi Tsenov, Valeri Mladenov
Abstract—The primary objective of this case study is to introduce a testing methodology framework for assessing the effectiveness of different drinks and flavors through neuromarketing. The framework involves use of focus groups for evaluation of the neurophysiological reactions of participants to various stimuli and drawing statistically significant conclusions based on the results. In the last past decade, there has been an increase in the availability of affordable and easy-to-use biosensors that provides the possibility for affordable and easy to use real-time measurements of bio-signals, such as electroencephalography (EEG), galvanic skin response (GSR), and pulse. Also, with the recent ongoing increase of the computational capabilities the signal processing and analysis of the recorded biometric data can also be processed in real-time with relatively affordable and mobile hardware. By analyzing these signals and identifying an informative features that are derived statistically from group averaging and independent of individual user’s deviations, it is possible to classify and then apply the measured brainwaves and other biometrics data in various applications. In this paper, the authors apply data from voltage potential electrodes EEG brainwaves, along with GSR and pulse in order to recognize the level of response variation among a test focus group, in order to determine their aggregated preferences on test sample drinks.
Keywords—biosignal feedback, electroencephalography, feature extraction, neuromarketing, signal processing

4003:
A Simplified Analog Neuron Model with Modified Memristor-based Positive Synaptic Weights
Stoyan Kirilov, Georgi Tsenov, Valeri Mladenov
Abstract — In this paper an analog circuit model for modified artificial neuron with memristor-based synapses is proposed. In this implementation each synaptic weight is realized by only one metal-oxide memristor, and this providing a vastly reduced circuit complexity. The summing and scaling device operation implementation is based on op-amp and a memristor. The activation function is realized by a simple circuit with CMOS transistors. The operation of the proposed neuron is analyzed both analytically and in LTSPICE environment and the derived results are compared and verified. The presented memristor-based neuron is a step for design of more complex neural networks with memristors.
Keywords — memristor, neural networks, synaptic weights, modeling and simulation, circuits and systems
4004:
Development of Advanced Control Strategy Based on Soft Actor-Critic Algorithm
Michal Hlavatý, Alena Kozáková
Abstract—The paper demonstrates the differences between two reinforcement learning methods – the Actor-Critic (AC) and the Soft Actor-Critic (SAC) applied for the benchmark pendulum and double pendulum control problems. The advantages and disadvantages of both methods as well as proposed modifications of the algorithms are discussed. The neural network has been developed using the Python and Tensorflow 2 libraries, and PyBullet Envs was used for environment simulations the case study.
Keywords— machine learning, reinforcement learning, neural networks, actor-critic, soft actor-critic, control, pendulum, double pendulum
4005:
Distributed Machine Learning through Transceiver Competitive Connectivity of Remote Computing Systems
Danail Slavov, Vladimir Hristov, Anastasiya Slavova
Abstract—An architecture is built for sequential data transmission among individual machines participating in a decentralized distributed machine learning of a convolutional neural network for image classification. Simultaneously running multiple replicas of a machine learning task using TCP communication protocol has been successfully implemented. A unification of the final result of each epoch until the end of the training session is performed on each of the participating machines in the training process. A working image classification model is obtained on all machines involved in the process. The results from the experiments performed herein show a 63% improvement in training speed for the fastest configuration with only a 3% drop in accuracy compared to the most accurate one. The study therefore confirms that the use of distributed learning systems can have significant benefits for the AI researchers and practitioners.
Keywords—machine learning, distributed algorithms, PyTorch, data parallelism, convolutional neural network
4006:
H∞ Reduced Order Control of a Thermal Plant
Bozhidar Rakov, Georgi Ruzhekov
Abstract — The current paper investigates the possibility of H∞ controller order reduction for a simple dynamic plants, which are predominand in the industry. A design procedure is carried out for a thermal plant model, received by system identification. Two full order controllers with integral part are obtained having series and parallel form. A mixed S/KS sensitivity structure is employed. Using Hankel singular values additional 5 reduced order controllers are obtained. Finally based on reduced plant model 4 new controllers are obtained. All control algorithms are implemented in digital form based on discrete state space equations, on a Programmable Logic Controller. Simple programs are developed for matrix addition and multiplication based on user defined PLC type. Additionally some programs are implemented for matrix transfer between a MATLAB client and SCADA OPC server. Experiments are conducted with a real plant. All eleven closed loop systems are compared by the singular values of their sensitivity functions. Performance characteristics are calculated and arranged on table. For the plant employed, the results show strong possibility of order reduction without significant loss of performance. The parallel form controller makes it possible to reduce the controller further that the series form.
Keywords— Hankel Singular values, H∞ theory, reduced order controller, PLC
4007:
H∞ Multivariable PID Control of a Thermal Plant
Bozhidar Rakov, Georgi Ruzhekov
Abstract — The current paper investigates the possibility to design a discrete multivariable H∞ PID controller using the MATLAB hinfstruct() function. Two PIDs are designed. They are compared against a full order H∞ controller and a decentralized PID with a static decoupling matrix. Frequency response of the singular values of the closed loop sensitivity functions are calculated and compared. All controllers are tested on a laboratory model, representing a thermal plant. The control algorithms are implemented on a Programmable Logic Controller. The elements of the multivariable PIDs are implemented using velocity form for simpler anti wind-up mechanism realization. The parameters are calculated by MATLAB and then transferred using OPC connection to a SCADA system. Two sets of experiments are conducted with the real system, testing two different sets of reference signal changes. Based on known performance criteria, a comparison is made. The results show the viability of a fixed order, fixed structure controller designed based on H∞ theory.
Keywords — H∞, fixed structure, multivariable, PID, PLC
4008:
Algorithm for Predicting Daily Volatility of FOREX Markets
Alexander Hotmar
Abstract— The paper addresses the issue of predicting daily volatility in FOREX markets based on past market movements. The method used is based on the combined use of two volatility indicators - day of the week and market trend. The studies were conducted over a large time period and multiple swaths to obtain a relatively universal estimation method and to avoid random influences due to the limited nature of the data samples. Combining the two indicators allows for partial compensation for random variation in each of them separately. The end result is a very good accuracy of the volatility estimate with an error of about 30%. This is a very good achievement, especially considering the use of only past values for the movement of the markets.
Keywords— FOREX, volatility, prediction, forecasting, daily
4009:
Simulating a Рobot’s Movement Using Game/Physics Engine
Aykut Ismailov, Vladimir Hristov
Abstract— In this paper a method to simulate robot/manipulator movement via the use of physics engine or game engine is proposed. A proof of concept implementation for a robot in 2-dimensional space is given. For the proof of concept application the Python programming language is used since it has a wide variety of libraries implementing the needed subsystems. The application allows realistic movement of the robot model and the modularity of the presented application allows for further development, which can unlock the ability for training artificial intelligence models for various tasks.
Keywords— robot, manipulator, simulation, physics engine, game engine
4010:
A Model-Free Method Based on Kautz Functions for Dynamic Measurement Improvement
Tsonyo Slavov, Miroslava Baraharska, Ivan Markovsky
Abstract— This paper proposes a model-free method for dynamic measurements improvement, in the case of time-varying measurement quantity. As an example, the oscillating mass measurement process is regarded. The proposed method uses a representation of measurement dynamics by the Kautz orthogonal model that is in contrast with the conventional model –free method, where the measurement dynamics is modelled by a specific type of finite impulse response model. In previous work, the authors developed method for dynamics measurement improvement, based on the orthogonal Laquerre model, which is appropriate when considering well-damped processes, whereas the Kautz model structures are suitable for oscillating processes, such as the observed one. Utilization of the Kautz model leads to estimating of a low order measurement model, which additionally simplifies the real time implementation of the suggested method. To estimate the time-varying measurement model a modified regression model is developed. Its parameters are estimated in real time by the recursive least squares method for time-varying models. The method adapts to parameter changes by keeping constant trace of the covariance matrix.
Keywords—Kautz series expansion, orthonormal model, recursive least squares with constant trace of covariance matrix, model-free method for dynamic measurement improvement
4011:
Non-Volatile Memories Based on Memristors
Stoyan Kirilov
Abstract — In this paper, a modified metal-oxide memristor model that is relatively simple, efficient, and improved is presented and applied for analysis of non-volatile and passive memory crossbars. A corresponding LTSPICE memristive library model is developed and applied for the studies. The suggested memristor model demonstrates a good performance even at high-frequency signals, effectively capturing the key properties of metal-oxide memristor elements. Furthermore, its effectiveness for operation in complex memristor-based circuits and devices is verified.
Keywords — metal-oxide memristors; enhanced memristor model; nonlinear ionic dopant drift; LTSPICE memristor library model; memristor-based memory crossbars
4013:
Methodology for Joint Work of Industrial Robots and Processing Centers
Metody Georgiev, Kostadin Milanov, Dilyana Gospodinova
Abstract — This article describes the method and steps for developing a robotic system based on an industrial robot-manipulator, a conveyor belt carrying polymer parts for mechanical processing, driven by a three-phase motor with appropriate frequency control, three single-phase processing centers (universal motors) for mechanically processing of small details, controlled by a three-phase frequency converter, and switching and safety equipment. The control of single-phase universal motors using a three-phase frequency converter and the use of chokes to reduce excessive voltage fluctuations supplied to the universal motors were the two main issues that needed to be resolved in order to implement the robotic system. To realize a correct and reliable operation of the entire system, it is crucial to realize robustness of the entire system.
Keywords — robot-manipulator, 3 phase frequency converter, single-phase universal machine
4015:
Robust Control of a Technological Module
Vesela Karlova-Sergieva, Boris Grasiani, Alexander Marinchev
Abstract— The proposed QFT algorithm guarantees steady state and stability in the whole working area of the control plant. The step-by-step application of QFT for control plant with significant parametric uncertainty resulting from nonlinear effects under different loads of the working tool is recommended as the proposed QFT algorithm guarantees steady state and stability in the whole working area of the control plant. The resulting control algorithm provides constant rotational speed of the working tool when its load changes.
Keywords— QFT Controller, Transient and Robust Performance, Turning, Milling, Grinding, Drilling
4016:
Application of Advanced Machine Learning Algorithms for Early Detection of Mild Cognitive Impairment and Alzheimer’s Disease
Zuzana Képešiová, Štefan Kozák, Eugen Ružický, Alfréd Zimmermann, Richard Malaschitz
Abstract—Alzheimer's disease (AD) is an incurable neurodegenerative disease that primarily affects the mental and linguistic abilities of the sufferer. The disease affects to a large extent not only the quality of life of the patients themselves, but also of their immediate environment. Diagnosing AD is a complex process that requires multiple resources, facilities and, above all, qualified experts. The whole process is time-consuming. In the early stages, such as mild cognitive impairment (MCI), it is difficult to diagnose the disease because of the very mild manifestations of the disease. To help medical specialists and researchers detect MCI and AD disease, we have developed a machine learning based approach. The patient is asked to describe in spoken word the offered images, either in a single word or a sentence description. This description is captured as an audio recording in a mobile app. The dataset of such recordings consists of more than 1500 recordings of different individuals, with approximately 9.49% of them having been diagnosed with MCI or AD. Due to the imbalance of the dataset, we used metrics such as Mathews correlation coefficient (MCC), sensitivity-specificity and ROC curve to evaluate the used machine learning algorithms. The analysis resulted in a machine learning model with 79.07% (54.35%-87.4%) MCC.
Keywords—machine learning; medical voice recordings; Alzheimer’s disease; Mild cognivie impariment; early detection
4018:
Novel approach for passive mixing in micorlfuidics utilizing porous PDMS sponge
Emil Grigorov, Boris Kirov, Jordan A. Denev, Vassil Galabov
Abstract—Mixing in microfluidic devices has been always a significant challenge due to the small channel dimensions and the strongly laminar flow character of the fluids. However, mixing is an essential process in various biological, chemical and pharma processes. This paper introduces a new approach to passive mixing in microfluidic devices, utilizing a quick and cheap way with a sugar cube as prototype to produce a) porous medium. The authors present experimental and numerical analyses of the method, showing that the efficiency of mixing decreases as the volume flow rates of the fluids increase for flows studied with Re in the range 0,028-0,26. The largest porous mediumtested achieved a maximum mixing efficiency of 98%, suggesting that larger geometries may yield even higher efficiencies. A decrease in mixing efficiency along the length of the sponge with increasing volume flow rates was also observed (experimentally and numerically), possibly due to increased backflow. Despite the complexity of the Polydimethylsiloxan (PDMS) sponge, numerical analysis indicates a linear relationship between pressure drop and Reynolds number.
Keywords - microfluidics, mixing, porosity, PDMS sponge
4019:
Servo Drive Control System
Kamen Hristov, Vladimir Hristov
Abstract— In the present article, a modern drive system for research the characteristics of servo drives is presented. The drive system is designed for researching different operating modes of a servo drive. The drive system is intended for implementation in the learning process at the Faculty of Automation. Various types of exercises have been developed for students to explore. Servo drives are electric drives for automatic control of torque, position, speed and possibly its derivatives using feedback. Feedback differentiates servo systems from other types of electric drives. In the case of servomechanisms, the command (coordinates, speed, moment, etc.) is fed to the input of the servo amplifier, which determines the current value, compares it with the set value and produces a control effect reducing their difference. With good reason, servo drives are accepted as the most high-tech electric drives.
Keywords— servo, vector control, position control, torque control
4020:
Stability Analysis of Electronic Circuits Using LTSPICE
Elissaveta Gadjeva, Ilona Iatcheva, Nikolina Petkova

Abstract—An approach to computer-aided stability investigation of electronic circuits using the general-purpose circuit simulator LTSPICE is proposed. The stability is analysed in the frequency domain according to the Nyquist criterion, K and  _criteria. The influence of parameter variation and design tolerances on stability characteristics is investigated. The computer implementation is based on pre-defined functions in the graphical viewer. Examples are presented to illustrate the developed procedure for stability investigation.
Keywords— stability study, parametric analysis, tolerances, CAD programs

4023:
An Implementation of Self-Organizing Multi-Stage Selection Procedure
Angel Marchev
Abstract— This research article presents an implementation of a self-organizing multi-stage selection procedure, a method that has its roots in the work of Alan Turing, Alexey Ivakhnenko, and Angel Marchev. The multi-stage selection procedure is applied to simultaneous structural and parametric identification of complex systems. The current implementation employs linear regression, a non-linear cross function, and is adaptable for various regression and classification problems. A MATLAB software implementation is demonstrated, and the method's advantages are compared with other approaches. A computational example is provided to illustrate the procedure's effectiveness.
Keywords — self-organization, Multi-stage selection procedure, A.I., MATLAB
4099:
Determination of Limit of Detection in Electrochemical Biosensors
Antonia Pandelova, Nikolay Stoyanov, Bozhidar Dzhudzhev
Abstract — one of the important parameters determining the quality of an electrochemical biosensor is the limit of detection (LOD). A definition of this parameter is given by the International Union of Pure and Applied Chemistry (IUPAC). Different approaches are known for calculating the LOD in different types of biosensors, each of which comes with advantages, disadvantages and limitations. Three methods are mainly applied to determine LOD from experimental data - from low concentration standard deviations, from a calibration curve, and from instrument resolution. The calibration curve method was used to determine the LOD in amperometric electrochemical biosensors. Experimental data from really developed electrochemical biosensors – plant-tissue and enzyme-based - were used to perform the calculations. The resulting values for sensitivity and LOD are b=2.97 mM/mgO2/l, SLOD=0.206 mM for plant-tissue biosensor and b=3.1 mM/mgO2/l, SLOD=0.176 mM for enzyme-based biosensor. The calibration curve method can be used as a useful tool for preliminary indicative quality analysis of a amperometric electrochemical biosensor.
Keywords — LOD, electrochemical biosensor, calibration curve

Last changed on 17.06.2023, 17:01:01
This page will be accessible until 01.12.2045

Other conferences/journals in the Conference/Journal Management System

Conference Management System © Technical University of Sofia 2012 - 2024.
Design by Assoc. Prof. Rossen Radonov, PhD, MEng.