Abstract: In recent years, Industry 4.0 has promoted enhanced integration of value chain participants and their production processes, aiming to improve their efficiency and effectiveness. In this paper, we address transparency and data privacy issues that occur with the introduction of a high level horizontal integration. We propose a use of private, permissioned Distributed Ledger Technology (DLT) systems for overcoming these issues and facilitating the integration of business processes with end-to-end engineering spanning across the entire value chain. We also present a tool for the automatic generation of smart contracts based on the production process specification. Generated smart contracts are used for transparent, secure and scalable monitoring of the production process executions.

Abstract: In recent years there has been a technological evolution that has provided the integration between different areas of knowledge, an example of this is the integration between Dentistry and Engineering in order to find new solutions to improve the surgical process of dental implants. This work proposes a conceptual method of reasoning to determine the type of dental implant based on the tomographic image where it will support the surgeon in determining the best dental implant options, depending on the characteristics of each patient. This system will analyse tomographic images through the bone structure, density and the load that implant will support, identifying the most appropriate models for the patient. The system creates a three-dimensional model of the dental arch and interacts with the dentist, leaving him free to choose among the selected ones the most appropriate implant. This system is a planning tool that assists the dentist during the preoperative period and in his / her decision making. Main contributions of the article are: i) design and development of a computational reasoning tool that supports the dental implant process; ii) interactivity in the development of surgical planning.

Abstract: Latest enhancements in IoT devices and in communication technologies, has brought new ideas that are capable of providing advanced sensing of the surrounding environment. On the other hand, average life expectancy has grown, resulting in a considerable increase in the number of elderly people. Consequently, there is a constant search for new solutions, to support an Active and Assisted Living (AAL) of these people. This paper aims to propose a solution to help in knowing the location of IoT devices that could be helping these people. The proposed solution takes into consideration the risk factors of the target persons, at any given time and as well as the technical constraints of the device, such as available power and communications. Thus, a profile-based decision is taken autonomously either by the device or its integrated system to ensure the use of the best geolocation technology in each situation.

Abstract: In this paper, the focus of the research is on the comprehensive flow-based anomaly detection architecture which is based on the joint use of the entropy calculation and machine learning algorithms, and its enhancement with time-series techniques. The proposed solution is evaluated with the modified CTU-13 dataset, which includes instances of normal, background and botnet traffic. The analysis encompasses a range of unsupervised machine learning algorithms, time-series and entropy threshold analysis with different configuration parameters.

Abstract: After a decade from its advent in 2008, the blockchain technology has dived deeply into mainstream research. The success of Bitcoin, used for storing transactions in blocks and thus maintaining a distributed ledger of who owns how many units of cryptocurrency, created high expectations for the possible usage of blockchain technology in the industry and the public sector. In order to step out from its limited use for finance sector and digital money, a new type of blockchains are emerging. The so-called smart contracts cast a light on possible usages of this technology in e-government, with the purpose to bring more transparency in e-government processes. Using the available features and data structures of the Smart contract it is possible to develop a solution for checking validity and assuring immutability of public documents.

Abstract: Big streaming data visualization has an important role in data exploration and mining, information retrieval and data analysis as it assists humans in analyzing and comprehending large volumes of streaming data. Systems for big stream data visualization must be able to handle massive amounts of data in a timely manner and provide an interactive and comprehensive graphical presentation. This paper describes a unified architecture for big streaming data visualization and lists the most popular libraries and platforms for its implementation. The described architecture supports all of the key characteristics of big streaming data: their volume, velocity, veracity and variety. Furthermore, we create a set of applications to demonstrate the viability of the described architecture, and to test some of the listed technologies.

Abstract: Vehicle tracking has become a common practice for commercial and traffic monitoring purposes. The problem of tracking the vehicle arises in the situation when there are more than one vehicle in the same place and the order of movement is important. Also, the problem is in places where there is a high frequency of vehicles such as parking spaces or public transport stations. The authors solve the problem through the application of different wireless technologies in monitoring the position of vehicles and their distance from the station, by automating the entry and referral of vehicles to the appropriate platform. This paper proposes an Internet of Things solution for controlling the entry and routing of a vehicle through reliability, security, and low cost. The authors applies a solution based on sensors, microcontrollers and wireless access control using radio communication, which is paired with infrared technology.

Abstract: Predicting when will a customer buy some product is a valuable information for any vendor. This information can be used for direct advertisement to customer or production/procurement planning. In this paper we focused on predicting whether there will be a purchase in the following seven days for a specific customer-product pair using LSTM neural networks. Considering that the data used are purchase records from a medical supply company, this research pertains to a B2B purchase scenario which has different characteristic than B2C purchases. Due to the irregular nature of purchase data, some data transformation was necessary before training the LSTM networks. Trained networks were evaluated on multiple datasets differing on the lowest number of purchases per customer-product pair. Results were compared to purchase prediction results from an earlier research based on the same dataset, but using ARIMA, a statistical method. Compared to ARIMA, LSTM networks’ predictions have lower precision, but significantly higher recall. Additionally, both precision and recall consistently rise when the dataset has a higher number of purchases per customer-product pair.

Abstract: In urban areas the most efficient, environment friendly and cost-effective method for supplying heat to buildings is by district heating systems (DHS). However, by optimizing the production of heating energy in DHS further improvement of the efficiency, operation cost reduction and raise of the profitability can be achieved. This goal cannot be met without a detailed analysis and adjustment of the heat supply according to the user demands. Accurate heat demand (especially peak load) is very hard to predict and Artificial Intelligence techniques like LSTM neural networks and DNN are most commonly used for development of Data Driven Models (DDM) for this prediction. In this paper various methods for heat load prediction applicable in District Heating System of Faculty of Mechanical Engineering in Nis (FMEDH) will be considered where Knowledge Based Models (KBM) and DDM for heat demand prediction will be analyzed while the goal that goes beyond State of the Art is development of Hybrid KBM/DDM model. Prediction models can further be used by the operators as a support for District Heat Plant Control as well as by the consumers as a decision making support for heat cost reduction if several heating sources are available (e.g. DHS, Solar energy, Heat pump, Air conditioning system etc.)

Abstract: The reconstruction of the missing part of the human mandible is a significant challenge in orthodontics and surgery, especially when the shape of the missing part is not known prior to operation. Creation of geometrical model of specific patient is being become an expected treatment. The one of the greatest challenges within this procedure is related to efficient designing the geometry and anatomy of a particular human bone. In this paper, method of the reconstruction of missing part the human mandible is presented. The reconstruction was performed by using the parametric model based on Method of Anatomical Features (MAF). The 3D parametric model is defined as point cloud, where coordinates of points are described by parametric functions, whose components are morphometric parameters (dimensions which are read on medical images). The main benefit of this model is ability to adapt model to the specific patient, by applying patient specific values of morphometric parameters. The resulted model may be used for preoperative planning and for surgical simulation. The main idea of the research was to present patient-specific reconstruction method which will provide the surgeon more control over the treatment, decreases the risk of surgical errors and reduces the operating time.

Abstract: The Internet of Military Things and the Internet of Battlefield Things (IoMT/IoBT) perform more different activities such as data collection, communication, data processing, command execution, data transmission. It is essential that subsequent activity is based on a decision made on the basis of authentic source data from a dynamic and unpredictable military environment. For that purpose, the authors proposed an authentication scheme based on PKI, digital signature and asymmetric cryptography. The proposed scheme does not require the establishment of an authentication channel for the authentication of the parties in the communication. Authentication is conducted on the IoMT/IoBT side after receiving the data. In addition to authentication, the schema provides data integrity and non-repudiation parties in transaction.

Abstract: In recent years, we have witnessed an increase in interconnected devices which collect and process data, and are used in a plethora of situations, giving rise to the Internet of Things (IoT). The measurements collected from these sensing devices - which in general may come from different manufacturers - must be aggregated, formatted, and processed together in order to provide integrated data management, and still presents a challenge to academia and industry alike. One of the critical applications of IoT is in healthcare management, where a strong motivation in government regulation should exist to secure protected health data transmission. A promising solution is integrating IoT with the still novel Blockchain technology, which is based on overcoming the lack of trust with its immutability and transparency properties. In this paper, we analyze state-of-the-art healthcarerelated Blockchain solutions with IoT (scientific papers, startup whitepapers, and readily available commercial solutions), identifying their similarities and differences, as well as possible bottlenecks. We hence provide an extended Blockchain/IoT healthcare model, based on multiple usecases which incorporates previously identified shortcomings, in order to build a basis for a trusted, reliable and secure Blockchain- and IoT-based healthcare system

Abstract: In recent years, we have witnessed the lack of control over our personal identities e.g. our birth dates, home addresses, and other personal data, that can be now easily accessed through search engines. Concerns also exist about more sensitive personal data, such as social security numbers, bank account information or health related data. One cannot be sure anymore which of one’s personal data is being shared through social networks or institutional systems, or even being modified without their knowledge. With all the different technologies that are in use, data often ends up being shared, modified or misplaced far more than one can realize. The aim of this work is to investigate how Blockchain technology can improve current, centralized data security solutions, with emphasis on medical records. In addition, we investigate how Blockchain can help keep data safe and in our control, following the regulations of various standards such as GDPR or, for medical records, HIPAA in the USA.

Abstract: K-means is one of the most important clustering algorithms, but it does introduce a risk of privacy disclosure in the clustering process. One approach to solving this problem is by applying differential privacy to K-means clustering algorithm to effectively prevent privacy disclosure. Increasing amounts of information generated in big data processing scenarios make clustering a challenging task. In order to deal with the problem, various approaches to the parallelization of clustering algorithms have been attempted. This paper presents an implementation of a differentially private k-means clustering algorithm that uses -differential privacy, based on the COMPSs framework for parallel computing. The experimental results show that the proposed implementation scales well and can be used to efficiently process large datasets using high-performance computing equipment.

Abstract: Simulation-based optimization integrates optimization techniques into a simulation analysis. Because of the complexity of the simulation, the objective function becomes difficult and expensive for an evaluation. In our example, the simulation considers structural analysis using the finite element method (FEM), the estimation of the welds number, the costs associated with the welding process, and the simplicity of the assembly process. To overcome the computing complexity, we are using OSICE, a comprehensive, cost-effective, and easy-to-use HPC/Grid and Cloud-based optimization service for solving large-scale optimization problems using parallel evolutionary algorithms. For this particular Application Experiment, after the initial definition of technical requirements, PAK input/output interfaces were redesigned to fit OSICE and CloudiFacturing requirements.

Abstract: This paper presents developing wrapper scripts for automating Finite Element Method (FEM) analysis on GNU/Linux servers. The purpose of these scripts is to edit data in ASCII files that are inputs for FEM solver and to call FEM solver which performs the analysis. Input files consist of geometry model, material parameters, loads, constraints, time step definitions and other data. After the long-lasting calculations, based on the stress results, material parameters in input files are updated and the analysis is restarted. This loop is repeated until the analysis predicts structure failure and for each pass safety factor is calculated. These scripts are also used to extract certain element groups, combine file sections and adjust output file for post-processing. Although Bash is very versatile when it comes to text manipulation it was necessary to augment it with Python programing language in order to achieve required functionality, primarily for fitting material parameters needed for next calculation. Repetitive, tedious work that an engineer needs to perform is greatly reduced, utilization of server time is improved, and this solution can be used for further development, for example, an inclusion of optimization, on which will focus in the future work.

Abstract: The basic tenet of the paper is related to the improvement of the implementation of business intelligence systems, which are important for deciding on a particular problem, and in order to achieve the highest level decision-making benefits. The improvement must be appropriate to the development of the BI system under conditions where there is already prior knowledge of the problem. The results obtained should show that the business intelligence system, enhanced by the application of R software package, can improve decision support.

Abstract: Network coding is a promising technique that improves network throughput and provides high reliability. In this paper we propose an application of quasigroups in network coding and give several advantages of proposed algorithm over linear network coding.

Abstract: The problem that we tried to solve was creating a system designed for young parents to provide information and assistance during the raising a newborn. Overcoming the obstacles and responsibilities of parenting is especially difficult for first-born parents. The large number of different often-contradictory information received from the environment, as well as the lack of adequate answers to many questions, is a constant problem. The amount of information available on the web can put an additional burden on parents. Currently available applications, as well as web portals, provide a wealth of information in textual form, which leads parents to read loads of text that very often makes no meaningful point. The available applications and web portals do not offer any interactivity that would allow parents to have a retrospective of the meal, the examinations performed, nor the ability to monitor the development of the infant and get suggestions from the system based on a review of all available infant parameters. The system proposal includes applications for monitoring the development of a child from birth onwards. The benefits of such a system are the ability to input relevant data from parents. Based on these data, system can analyze and inform parents when it is time for a meal, how many meals a baby should have, to suggest introducing a certain type of food based on what has already been introduced, to monitor child development, completed and future vaccination as well as exercises for proper development.

Abstract: JavaScript is now one of the most popular programming languages in the software development industry, and as it is growing, a lot of new JavaScript frameworks have entered the market. JavaScript frameworks enable easier and faster front-end development, but with a wide variety of frameworks available, developers can often find themselves in a dilemma choosing which framework is the best for them. This paper focuses on the three most prominent frameworks (Angular, React and Vue) and performs a detailed comparison based on specific applications that were made and a survey that was conducted. The parameters by which the comparison was made are: popularity, performance, community support, learning curve, migrations and flexibility. As a result of this comparison, a manual was created to help both beginners as well as experienced developers, to choose the right JavaScript framework, depending on their needs and abilities.