737 papers found .

361. Handling structured data in the Alfresco system

ICIST 2011 Proceedings, 78-82
Sladić Goran, Gostojić Stevan, Milosavljević Branko, Konjović Zora
Abstract: This paper describes an extension of the Alfresco system which provides retrieval of data from other relational databases. Retrieved information can be used as metadata for folders, documents and business processes within Alfresco. The proposed solution is compliant with the Alfresco mechanism for defining content models and therefore can provide a data source for different Alfresco installations.

362. Simulation Environment for Testing Autonomous Vehicles

ICIST 2018 Proceedings Part I, 6-9
Komavec Mojca, Balažic Marko, Sodnik Jaka
Abstract: The focus of automotive and software industry in the recent years has been in the development of algorithms for autonomous vehicles, whose input depends on various sensor data. Such algorithms which are mostly based on machine learning techniques need a vast amount of data to be trained successfully. Since obtaining data in a real world can be expensive and time-consuming job, one can obtain and train algorithms also in a simulation environment. Vehicles in a simulation also perceive the environment through a set of sensors whose output should replicate the output of real-world sensors as much as possible. This paper describes an example of such professional driving simulation environment for testing self-driving vehicles, gives an overview of a simple autonomous driving algorithm, which is partly based on sensor input, and concludes with an evaluation of our algorithm in comparison with the human-driving data. The results show that autonomous algorithm performed much better than human drivers.

363. Demonstrating Enterprise System Security Using an Asset-Centric Security Assurance Framework

ICIST 2018 Proceedings Part I, 16-20
Luburić Nikola, Sladić Goran, Milosavljević Branko, Kaplar Aleksandar
Abstract: Enterprise systems often require a high level of security. While securing such systems is a challenge in and of itself, proving that a system is sufficiently secure is an additional problem which is rarely discussed. This paper presents a security assurance framework that can be used to prove that an information system is reasonably secured. The framework is structured around data assets, their security goals, and their flow throughout a system. We present the domain model for our framework and describe how the asset inventory, data flow diagrams, and security assurance cases are used by it. Finally, we demonstrate how the framework produces a scalable body of evidence that can be used to demonstrate the exercise of due care and due diligence, mitigating some of the issues that arise from expert judgment-based approaches.

364. Two-step process for secure registration of nodes in IoT systems

ICIST 2018 Proceedings Part I, 28-31
Stojkov Milan, Simić Miloš, Sladić Goran, Milosavljević Branko
Abstract: Edge computing is recognized as an architecture which is well suited for the integration with the Internet of Things. A new layer of devices, also called nodes, between cloud services and end-users, can provide new services with fast response and great quality. On the other hand, new layer devices introduce more architectural and security challenges which were not yet considered. In this paper, the main focus is directed towards security issues, especially those concerning safe (re)authentication of the nodes in the already established IoT system. We defined a two-step process for a secure registration of the nodes in IoT system that includes a solution for protection against the physical replacement of the nodes with corrupted ones. Our approach uses public key cryptography based on ECCDH algorithm and communication between nodes, gateways and authentication services in the cloud.
Abstract: Smartphone sensors are being increasingly used in various applications in healthcare and sport. The performance of sensors varies among different smartphone models. A publicly accessible resource containing real-life-situation describing smartphone sensor parameters could be of great help for developers of applications in healthcare and sport. Therefore we have designed and implemented a participatory sensing application for measuring, recording, and analyzing smartphone sensor parameters. The application database includes sensor parameters of a number of different smartphone models. The database is offering information on several statistical parameters of the measured smartphone sensors and insights into their performance. The database of smartphone sensor parameters may prove particularly useful for developers of applications in healthcare and sport where inertial sensors are used. Index TermsSmartphone inertial sensors, Participatory sensing, Applications in healthcare and sport.

366. A comparison of vehicles user interface modalities in terms of User Experience

ICIST 2018 Proceedings Part I, 40-43
Čegovnik Tomaž, Tartalja Igor, Sodnik Jaka
Abstract: In this paper we report on a user study which focuses on user experience when users interact with In-Vehicle Infotainment System (IVIS) through three different input modalities. We compared two relatively new approaches such as touchpad-based input and free hand gesture input to the widely used input interface based on buttons on the steering wheel. In order to evaluate the user experience provided by three different input interfaces of IVIS the User Experience Questionnaire (UEQ) was used. The results of the study are showing clear differences between different input devices favoring the conventional buttons on the steering wheel.

367. Improving Library Services using Business Intelligence

ICIST 2018 Proceedings Part I, 44-47
Tešendić Danijela, Boberić Krstićev Danijela
Abstract: Paper describes model of data warehouse and usage of OLAP tools in order to support decision making in library. Usage of Business Intelligence technologies is presented on case study of libraries which use library management system BISIS and have considerable need for making various statistical reports. Focal point of this paper is dedicated to modeling of data warehouse using data sources from system BISIS and taking into consideration user requirements regarding reports.

368. An Overview of IoT Readiness Assessment Methods

ICIST 2018 Proceedings Part I, 48-53
Stankovski Stevan, Ostojić Gordana, Arsenijević Dejan, Baranovski Igor, Oros Dragana
Abstract: In the next decade significant influence of IoT technologies over all areas of human society is expected. Most research show that IoT implementation is in early phase in majority of companies, IoT application is partial, at starting point, through pilot projects (HCL Global Systems’ “Global IoT Report 2017” reports that 6 out of 10 surveyed are in early phases of exploration and defining IoT strategy). Potential benefits and necessity of application are not in question, but ways how to implement IoT efficiently and effectively are not enough explored and known yet. In defining IoT strategy it is necessary to assess current organization state from IoT perspective, IoT readiness assessment methods provide structured approach to do that. A number of typical IoT readiness assessment methods are presented in this paper.

369. Discovering Human-like Driving Strategies with Learning and Optimization Approaches

ICIST 2018 Proceedings Part I, 59-64
Dovgan Erik, Bratko Ivan, Sodnik Jaka
Abstract: Human driving strategies can be obtained in various ways. This paper presents two artificial intelligence approaches: a learning approach that builds a human driving model based on driving data, and an optimization approach that searches for driving strategies by taking into account existing human driving models. The results show that the optimization approach can optimize several objectives in addition to human likeness thus producing a set of driving strategies with various tradeoffs between the objectives, while the learning approach aims at exactly reproducing the human driving, which results in a single driving strategy.

370. The Watermark Applications in Multimodal Biometric Identification

ICIST 2018 Proceedings Part I, 65-69
Veličković Zoran, Milivojević Zoran, Veličković Marko
Abstract: The need for personal identification in the modern world in all areas of social activity is growing. This paper presents the application of the watermark in multimodal biometric identification. In one biometric feature - a face photo is inserted in the form of a watermark, a second, encrypted biometric feature - a fingerprint. Unlike standard biometric identification algorithms that require access to a database with biometric features, the concept of multimodal biometric identification does not require access to the database. It has been shown that a high quality fingerprint can be extracted from the watermarked photo of person, which, together with the face image, is used in the process of multimodal identification. In the experimental part of the work, mean values of PSNRs of secured images of over 37dB were obtained, while the average value of PSNR watermarks was over 33dB, NC greater than 0.98 and SSIM about 0.96. Based on the obtained experimental results, it can be concluded that the proposed concept of multimodal biometric identification based on a watermark can be successfully applied in multimodal personal identification.

371. Assessing uncertainty in the hydroclimatic modelling chain by using hydroinformatics platforms with varying complexity

ICIST 2018 Proceedings Part I, 70-73
Stojković Milan, Obradović Dragana, Milivojević Nikola
Abstract: Uncertainty in hydrological modeling has played a significant role in assessing the impacts of climate change on water resources. The considerable uncertainty in the predicted streamflows is more likely lies in climatic modelling depending on the chosen Global Climate Models (GCMs) and Regional Climate Models (RCMs), choice of the bias correction method, the assumed initial and boundary conditions, the chosen greenhouse gas emission scenarios and scenarios of future socio-economic development. Also, there are uncertainties related to the hydrological modeling scheme caused by input hydrological data, hydrological model structures and parameterization of hydrological model. In this research, the uncertainties the projected streamflows due to climate change are addressed by the direct variance method since this technique is capable for assessing different sources of uncertainties nested into the hydroclimatic modelling chain. In the case of climatic modelling four RCMs are used to derive inputs for hydrological models at river basin spatial scale under different greenhouse gasses emission scenarios. Namely, the RCP 4.5 and 8.5 pathways are used for this purpose. For hydrological modelling, the Precipitation-Runoff Modeling System (PRSM) and the 3DNet Catch hydrological model are applied. The research is carried out for the Lim River basin, while the simulations covered the 2011-2040 time frame with respect to the bassline time period 1961-1990.

372. The Recommendations for Green IT Strategy Implementation in Public Sector

ICIST 2018 Proceedings Part I, 74-79
Bjeladinović Srđa, Asanović Marko, Gospić Nataša
Abstract: Adequate planning, design, introduction and use of information and communication technologies (ICT) can enable organizations to automate and speed up business processes, reduce the number of possible mistakes caused by human factors, and increase the availability of timely information. Although all of these can have a positive effect on the business, the emergence of negative effects while using modern technologies are not negligible (increased emissions of harmful gases, unplanned and uncontrolled use of non-renewable resources and energy sources, etc.). The mentioned problems caused by the use of ICT can be solved using different techniques, and one of the most popular is Green IT. By analyzing the available strategies for the implementation of Green IT, it has been established that there is room for expansion of the same, in order to more adequately meet the needs of the public sector in the territory of Serbia and Montenegro. In this paper, emphasis is placed on identifying the basic aspects of the strategy for the implementation of Green IT in the public sector, as well as providing a set of recommendations for each of these aspects. This paper will contribute to the popularization of Green IT in the public sector, and the recommendations given in it may serve as a starting point when the overall Green IT implementation strategy is created.

373. Framework for connected supply chain based on Internet of Things and cloud services

ICIST 2018 Proceedings Part I, 80-83
Stefanović Nenad, Milošević Danijela
Abstract: With the power to enable greater visibility and control, the use of Internet of Things (IoT) technology in the supply chain offers new possibilities to improve processes and operations. However, due to supply chain complexity, implementing IoT solutions effectively can be very challenging project, and it requires forethought, strategy, platform and models. In this paper, we discuss opportunities for employing IoT systems in supply chain management (SCM) along with literature review and state of the art background research. We introduce the six-step framework for implementing more agile and intelligent IoT connected SCM systems. Finally, based on the proposed framework, we present and-to-end solution for connected factory which utilizes various IoT, data management and analytical cloud services in order to enhance visibility and responsiveness, increase efficiency, lower costs, and improve decision making.

374. Formal Semantics in 3D Cadastre

ICIST 2018 Proceedings Part I, 84-89
Sladić Dubravka, Radulović Aleksandra, Govedarica Miro
Abstract: The paper presents LADM based Serbian cadastral domain model, as a base application schema for the ontology development. After that, rules for formalizing an application schema in OWL, defined in ISO 19150, have been followed to obtain standard based formal ontology for Serbian cadastral domain model.

375. Robust equipment design for a real-time coaching feedback support in watersports

ICIST 2018 Proceedings Part I, 90-93
Umek Anton, Dopsaj Milivoj, Kos Anton
Abstract: In watersports, the use of on-site technical equipment for coaching support is not very common. Coaches most frequently use only a stop-watch. Wearable sensors attached to the athlete or installed into and onto sport equipment offer valuable supplemental information to coach during the practice. A technical system providing real-time feedback can make coaching more efficient. The paper presents results of our field tests in swimming, canoeing and kayaking. All listed watersport disciplines have many very similar parameters important to coaches, such as stroke frequency, stroke duration, stroke symmetry, and others. We focus on the definition of demands for a measurement system that would best fit watersport professionals. Our final goal is the development of a real-time feedback system consisting of waterproof wireless devices with multiple sensors supported by user friendly application. Sport professionals should be able to operate the system without constant presence of trained and experienced technical personnel (engineers). Index Termswearable sensors, smart sport equipment, feedback system, watersports.

376. Emergence of new technologies in the designing and fabrication of patient-specific removable partial dentures (RPDs): A case study

ICIST 2018 Proceedings Part I, 94-99
Šokac Mario, Budak Igor, Santoši Željko, Kadvany Kristian, Puškar Tatjana, Vukelić Đorđe
Abstract: Application of 3D technologies brings new innovative tools that support design process of custom implants and dentures, more specifically related to the field of dentistry. These new tools, which allow easier and faster designing process of patient-specific removable partial dentures (RPDs), enables them to be widely accepted as a more preferable choice to use. In this paper, implementation of 3D technologies will be presented through a case study for designing and fabrication of patient-specific RPDs. The entire process will be presented and elaborated, starting from the designing stage of a 3D model of a patient-specific RPD using advanced tools, all the way to a verification stage using RP (Rapid Prototyping) technologies. This approach enabled a more convenient design and faster delivery time of a manufactured RPD, but also by using less expensive equipment to achieve this.

377. An Application of Blockchain Distributed Systems for Supply Chains in the Pharmaceutical Industry

ICIST 2018 Proceedings Part I, 100-104
Mirković Aleksa, Nenić Marina, Gajić Dušan, Terzić Branko, Luković Ivan
Abstract: This paper describes an application, built on top of the Hyperledger Fabric blockchain framework, for efficient and reliable support of the supply chain in the pharmaceutical industry. The presented application showcases how blockchain’s immutable ledger, peer-to-peer (P2P) network, consensus algorithms, and smart contracts all combine and synergize to create a secure, reliable, auditable, and cheaper way of performing business. The application is developed on the Proof of Concept (PoC) level, using the Hyperledger Composer framework. It enables communication with a private Hyperledger Fabric network via Create, Read, Update, and Delete (CRUD) operations. The developed PoC shows that the use of the Hyperledger Composer framework allows simple creation of business networks which include participants, assets, transactions, and relationships between these entities. Our research shows that Hyperledger Composer significantly reduces the amount of time needed to develop a Hyperledger Fabric blockchain-based software solution. However, for more customized approaches, the Hyperledger Fabric platform needs to be used directly on the infrastructure level. When using Fabric, smart contracts are written on the lower level of abstraction, using the Go programming language, instead of the high-level JavaScript code used in Composer. Plans for future work include implementing the solution directly on the Hyperledger Fabric platform, as well as extending the available functionalities of the application to cover an even wider set of use cases.

378. Serbian Legislation as a Network

ICIST 2018 Proceedings Part I, 105-108
Vidaković Dragan, Gostojić Stevan, Kovačević Aleksandar
Abstract: The recently introduced concept of a legislation network is a promising approach to addressing the challenging issues of tackling with and quantifying the complexity of the legislation. In this paper, we described the process of crawling and scraping Serbian legislation, applying natural language processing methods to detect references in legislation and applying network science methods (e.g. node centrality and community detections) to quantify network properties. The quantified network properties were attributed with corresponding legal interpretations. Keywordslegislation network; legislation; network science; natural language processing; computational legal studies;

379. A Comparison of Daily Human Activity Classification Techniques for Wearable Devices

ICIST 2018 Proceedings Part I, 109-111
Stančin Sara, Tomažič Sašo
Abstract: We present and compare some of the most common techniques used for classifying daily human activates that are applied to signals from wearable kinematic sensors. We investigate and compare different research results reported on this topic that focus on the most common human activities, i.e., walking, running, sitting, lying and climbing up stairs. Besides comparing the considered research results in the obvious terms of classification accuracy, we also compare them with respect to the final number of features considered and the length of the windows used for feature extraction. Both of the latter characteristics have an impact on computational efficiency and delay and so affect the suitability for real-time operation of the presented techniques.

380. Extensible Python Library for Managing Probabilistic Knowledge Structures

ICIST 2018 Proceedings Part I, 112-115
Vidaković Dragan, Segedinac Milan, Konjović Zora, Savić Goran
Abstract: In this paper we have proposed an extensible Python library aimed at utilization and managing probabilistic knowledge structures. A key part of the library is implemented focusing on the probabilistic knowledge structures application to educational domain. The library is structured as three Python modules aimed respectively at: (1) data representation and conversion, (2) basic modeling of local independency, and (3) gradation of a probabilistic knowledge structure. The library was verified against the requirements specification and the obtained models are validated by the chi-square test. The data set used for verification and validation is a subset of PISA testing data containing responses of 340 German students on 5 questions from mathematical literacy. Keywordsprobabilistic knowledge structures; knowledge spaces; parameter estimation;


Spread the word!



Eventiotic is a platform for provision of the unique digital services to scientific events' organizers, participants, and other interested communities.

Eventiotic repository facilitates universal and unrestricted access to the scientific events' online proceedings, for better visibility, more citations and more collaborations.