Abstract: In Minimally Invasive Surgery (MIS), surgeons need to acquire a specific set of skills, before carrying out a “real” operation. Training with the Laparoscopic Surgical Box-Trainer device helps in acquiring the needed skills for surgery residents which are traditionally not taught to them. Video recording of residents’ performance and computer-assisted surgical trainers for MIS provide valuable information for resident’s assessment. In this paper, we propose real-time detection and tracking of a multi-class of laparoscopic instruments for an intelligent box-trainer performance assessment system using SSD-ResNet50 V1 FPN architecture in TensorFlow backend. The dataset has been extracted from various laparoscopic box training videos. Using distance measurements and evaluation criteria constraints, we present an evaluation of the surgeon’s performance. Based on the experimental result, the trained model could identify each instrument at the score of 90% fidelity, in each location, within a region of interest. This research is a result of a partnership between the Department of Electrical and Computer Engineering and the Department of Surgery, of the Homer Stryker M.D. School of Medicine, at Western Michigan University.

Keywords: Intelligent Laparoscopic Surgical Box-Trainer; Laparoscopic Surgical Tool Tip Tracking; Fuzzy Logic-Based Performance Assessment System

Abstract: There are many suitable global optimization approaches to find the minimum value of an objective function. In this paper, the improvement of the GLOBAL Optimization Method is studied, which is based on stochastic clustering. Through its three main components, which are sampling, clustering, and local search the algorithm aims to find the global minimum of the objective function. Local search methods significantly influence the efficiency of the GLOBAL method. The efficiency of our proposal may be improved by dividing the system into modules and by creating new variants of both the local and line search methods. The main contribution of this work is to show the achievements of modularization and the efficiency of the new variants of both local and line search methods.

Keywords: GLOBAL; Optimization; Local search; Line search; Modular software

Abstract: One of the biggest problems nowadays in the online environment is the spreading of misinformation. Especially during a global pandemic, the most popular topics of fake news are related to coronavirus. Therefore, automatic detection of such news in the online media or social networks can help with the prevention of misinformation spreading. During the recent years, deep learning models proved to be very efficient in this task. However, the majority of the research focuses on the training of these models using publicly available data collections, mostly containing news articles written in the English language. As the spreading of fake news is a global phenomenon, it is also necessary to explore these approaches on the various local data sources. The work presented in this paper focuses on using the deep learning models for the automatic detection of fake news written in the Slovak language. We collected the data from multiple local online news sources related to the COVID-19 pandemic and used it to train and evaluate the various deep learning models. Thanks to the combination of bidirectional long-short-term memory network with one-dimensional convolutional layers, we achieved an average macro F1 score on an independent test set of 94%.

Keywords: natural language processing; deep learning; convolutional neural networks; fake news; COVID-19

Abstract: From a control perspective, the exact conditioning of systems with time-varying parameters is still a challenge. Many adaptive control algorithms (Adaptive Inverse Dynamics - AID, Model Reference Adaptive Controllers - MRAC, etc.) exist today. "Fixed-Point Iteration Methods" attempt to offer "alternative" control planning methods to circumvent the application of the Lyapunov function technique. The foundations of the method were developed in 2009. RFPT is an iterative control method, based on the fixed-point theorem of Stefan Banach proved in 1922 [1]. There is usually no specific suggestion for the choice of controller parameters, as the response function also depends on the approximation model parameter used and the actual behavior of the system under control. Adaptive RFPT presupposes strongly nonlinear system models in the first place, so in this case, thinking in frequency image and step inputs is not relevant (it is not advisable to conflict a nonlinear system with step inputs), so it does not have a tuning technique applicable to LTI models. However, there are a number of optimal search methods that can also be used to tune controllers (e.g., PIDs), e.g. the Genetic Algorithm. Using this method, I developed a possible autotuning process for adaptive RFPT controllers.

Keywords: Adaptive Control; Fixed Point Iteration-based Adaptive Control; Banach Space; Genetic algorithm; GA-based RFPT auto-tuning; Auto-tuning method; Control systems

Abstract: Images are affected by noise during their acquisition and transmission. Therefore, the denoising process is necessary to achieve higher quality images. However, both edges of the image and noise are characterized by high frequencies, loss of edge information may become unavoidable as a result of the denoising process. Thus, recovered, denoised images, become blurrier or less denoised. Therefore, a wavelet threshold denoising technique, based on edge detection, can be used to preserve more edge information and enhance the quality of the denoised image. In this paper, a novel image denoising method, based on wavelet thresholding by using Otsu's threshold, has been proposed and the clarity of the image which has been handled with this method is superior to that currently achieved by the other wavelet thresholds. The obtained results show that the proposed method, in this paper, provides better performance compared to commonly used wavelet image threshold denoising methods in terms of the visual quality of the denoised image. In addition, when the edge detection and denoising processes are combined, the deficiencies of the commonly used denoising methods are eliminated and a better denoising effect has been achieved.

Keywords: wavelet threshold; image denoising; wavelet edge detection; peak signal to noise ratio (PSNR)

Abstract: This work discusses the LMS (Learning Management System), as part of the University's Educational Portal and identifies the main hardware/software components of such an e-learning system. The conceptual model is presented, in the form of a queuing network. Analysis of the performance and reliability of the system, is used, as criteria. The Authors assessed the productivity of the EKTU e-learning system, based on the proposed model.

Keywords: learning management system; information and communication technologies; on-Line education; distance learning; educational environment; educational portal

Abstract: This paper considers the determination method of deformation location of lithium polymer batteries. Measurements are performed using the Digital Image Correlation (DIC) technique and the obtained results are sorted into a database as a function of the charge level. A statistically based algorithm is used to eliminate measurement errors and outliers. This paper adopts the Weighted Ordered Weighted Averaging (WOWA) operator-based 2D filtering method with the purpose of determining the critical regions of the cell. During the tests, several lithium polymer batteries of the same type but in different states are compared. Measurements on completely new and also on worn-out batteries are performed. The results support that the regions where greater deformation is expected during charging and discharging can be predicted. Results of investigations validate that the proposed approach is suitable for determining the critical deformation regions with high accuracy.

Keywords: battery swelling; battery testing; lithium polymer battery; GOM Atos; DIC measurement; WOWA

Abstract: The commercial success of businesses depends to a large extent on the ability to quantify the current situation at the desired point in time. It helps to make the right strategic development decisions and reveals weaknesses. The state of development of a company is a complex phenomenon, therefore, it can be described only by a certain number of indicators. They are multidimensional and of unequal importance, therefore, multiple-criteria methods are used to combine their values into one generalising quantity. They rely to a large extent on expert assessments to determine the weights of indicators, as well as the values of indicators that are difficult to formalise. An integral part of such assessments is the examination of the consistency of expert opinions. Existing methods for determining the level of consistency of expert assessment are intended to determine the importance of indicator weights, and it is not possible to determine the consistency of expert assessment of indicator values on their basis. This is because, when determining the importance of an indicator, the estimate of the importance of one indicator follows from the context of the importance of all other indicators, whereas in the second case, the value of each indicator is determined separately, i.e., it does not follow from the context of the values of the other indicators. In order to determine the consistency of the expert assessment of the values of indicators, it is necessary to calculate the actual and maximum possible level of uniformity or non-uniformity of the assessment. The consistency of the expert assessment will be demonstrated by the ratio of these values. The aim of the article is to propose and approve a methodology for determining the compatibility of expert assessment of the values of difficult-to-formalize indicators that increase the commercial success and competitiveness of business enterprises.

Keywords: business enterprises; multiple-criteria methods; compatibility of expert assessments

Abstract: The modeling problem is one of the important topics in engineering applications. In various applications, it is required to find a mathematical model to represent the relationship between output and the associated input variables. In this study, an approach on basis of a new deep learned type-3 (T3) fuzzy logic system (FLS) is introduced. The modeling of CO2 solubility on basis of temperature, molality of NaCl, and pressure is considered as an application. The monitoring of carbon dioxide (CO2) solubility in brine is one of the effective approaches in carbon capture and sequestration technique to reduce it in the atmosphere. A new hybrid learning method is presented to optimize the suggested model. The new adaptation laws are carry-out to tune the rule parameters and centers of membership functions (MFs). The values of horizontal slices and α- cuts are learned by the unscented Kalman filter (UKF). By the real-world experimental data sets, several statistical examinations, and comparison with conventional well-known fuzzy neural networks (NNs) and learning methods, the reliability and good performance of the suggested method are demonstrated. Also, the sensitivity of the input variables is analyzed by the use of the Sobol approach.

Keywords: Carbon dioxide solubility; Fuzzy logic systems; Learning algorithm; Estimation performance; Kalman filter

Abstract: Arc teeth semi-rolled cylindrical gears in the conditions of unbraced machine bodies have a higher load capacity, durability and reliability as well as the ability to compensate for the shaft axis twist angle by self-adjustment of one of the wheels compared to straight and helical teeth gears. In the article, the study object is the arc teeth semi-rolled cylindrical gear. The wheel arc teeth are cut using the single division method without generating with the cutting head, the generating surface of which is a straight circular cone. The gear arc tooth flank is an envelope of the wheel teeth flanks family at a given relative movement. The semi-rolled version of arc teeth cylindrical gears allows to significantly simplify the technological process of cutting wheels and producing gears with large gear ratios. Mathematical models of wheel and pinion arc teeth flanks forming processes have been built for a semi-rolled cylindrical gear. Dependences for calculating principal curvatures of the wheel and pinion arc teeth flanks have been obtained. An algorithm and a program for calculating the coordinates of the points of the active path of action in arc teeth meshing and principal relative curvatures at these points have been developed. The problem of determining the size of the contact pattern in the studied gear has been solved. Calculations to estimate variations in the position of the active paths of action and the sizes of the contact pattern with varying the wheel and pinion axes twist angle and variations of the principal relative curvatures in the longitudinal and profile directions of the arc teeth at the points of active paths of action have been performed.

Keywords: semi-rolled cylindrical gear; arc teeth; teeth flank curvatures; lines of action; principal relative curvatures; contact pattern

Abstract: In this paper a new numerical procedure is developed for calculating the inclination angle and deflection as initial conditions of the end points of statically determinate and indeterminate beams. The method is based on the topology comparison of simple (hinge-roller combination) supported beam and a resemblant cantilever beam. Assuming that the support reactions of the beam are active forces, the virtual displacements at the points of the reaction forces are calculated. Based on these values the inclination angle is calculated. Several examples are considered and the suggested in this paper, while the procedure is applied for various types of structures and loadings. The results, obtained by the suggested numerical procedure, are compared with analytical ones, and they are in good agreement.

Keywords: elastic curve; beams of variable cross section; initial guess for slope and deflection

Abstract: In Discrete-Event Systems (DES), deadlocks frequently occur. Flexible Manufacturing Systems (FMS) have the character of DES. Namely, FMS consist of many cooperating devices (like robots, machine tools, transport belts, etc.). Frequently, deadlocks occur because of insufficient resources. Petri Nets (PN) are often used to model FMS and to synthesize control for them. To deal with deadlocks, first of all, it is necessary to find and/or avoid them. There are several principal approaches for doing this - either by computing and analyzing the PN reachability tree (RT) or by finding PN model siphons. Then, in the former concept, the supervisor is synthesized by means of P-invariants of the PN model used, while in the latter concept the supervisor, based on siphons, is synthesized. In addition to these approaches, additional techniques can sometimes, be applied - e.g. a suitable utilization of added PN transitions.

Keywords: control; deadlocks; discrete-event systems; Petri nets; P-invariants; siphons

Abstract: Today, the widespread use of information and communication tools along with the developing technology has facilitated access to information. These developments have revealed the importance of data security. Many encryption algorithms have been developed to ensure secure data transfer. In this article, we have developed a new Genetic Encryption Algorithm (GEA) inspired by the DNA structure. The GEA is compared to a DES (Standard Encryption Algorithm), an AES (Advanced Encryption Algorithm) and a RSA encryption algorithm. A short evaluation is made, presenting the results, along with tables and graphs.

Keywords: Cryptology; Genetics; Encryption; Algorithm; Performance