Abstract: A procedure for modelling of mean monthly flow time series using records of the Great Morava River (Serbia) is presented in this paper. The main assumption of the conducted research is that a time series of monthly flow rates represents a stochastic process comprised of deterministic, stochastic and random components. The former component can be further decomposed into a composite trend and two periodic components. The deterministic component of a monthly flow time-series is assessed by spectral analysis, whereas its stochastic component is modelled using cross-correlation transfer functions, artificial neural networks and polynomial regression. The results suggest that the deterministic component can be expressed solely as a function of time, whereas the stochastic component changes as a nonlinear function of climatic factors (rainfall and temperature). The results infer a lower value of Kling-Gupta Efficiency in the case of transfer functions, whereas artificial neural networks and polynomial regression suggest a significantly better match between the observed and simulated values. It seems that transfer functions fail to capture high monthly flow rates, whereas the model based on polynomial regression reproduces high monthly flows much better because it is able to successfully capture a highly nonlinear relationship between the inputs and the output.

Abstract: Data gathered through decades of hydrological monitoring practice, often through elaborate measurement strategies is crucial for the process of decision making in electrical energy production. There is a substantial need for temporal and spatial integration of gathered data which is best achieved by accessible software solutions. We have developed hydro-information software system to handle data complexity through acquisition, organization, analysis, storage, and data presentation to end users. We have used modified server-client architecture by combining the advantages of non-structured databases with the power of modern web-development techniques. Our system is now installed and operational in hydro-electric power industry environment, and its users now have intuitive, flexible, and detailed access to all aspects of the gathered data as well as facilities to model, observe and forecast future data.

Abstract: For a long time, novel analytical models have been applied in modeling, prediction and monitoring of water losses and seepage in hydrotechnical objects (HO). Statistical models based on multiple linear regression (MLR) have been shown to be more or less successful for modeling processes in many domains, but have not been a common choice for modeling the above mentioned processes. The rapid improvement of sensor and data acquisition technologies enables the development of statistical models in basic and advanced MLR forms, such as hierarchical regression, stepwise regression (SWR), robust regression, and ridge regression. This paper presents a framework for the application of advanced statistical tools in modeling and evaluation of water losses and seepage in hydrotechnical objects, with the focus on stepwise regressions. The developed framework provides a platform for software generation of adequate regression models of seepage and water losses, considering both the number, and the type of regressors.

Abstract: A thermal numerical analysis of large infrastructural objects such as dams, bridges, tunnels, buildings, etc., requires the details about the structure geometry, loadings, boundary conditions and carefully determined material parameters. The material parameters obtained by an expert opinion or an experimental identification can vary from the parameters of specific real structure. This can give inadequate results and the significant difference between the measured data and computed results. To overcome this problem, it is necessary to develop and prescribe the methodology for material parameters calibration. In this paper, one possible approach is applied on example of gravity dam. The proposed methodology consists of: 1) the huge dam model reduction to substructures (lamellas) with the best quality of measured data, 2) the material parameters sensitivity analysis, 3) the calibration using intelligent methods (artificial neural networks, genetic algorithms, etc.) and 4) the verification by comparison of numerical analysis of the whole structure using the calibrated parameters and the available measured data.

Abstract: Paper introduces model for adaptive e-business continuity management (e-BCM) in insurance sector, which is adjustable to changes in the business environment of the organization. The research is focused on improvements to the establishment of effective business continuity management in insurance companies that use modern e-business technologies: the Internet, mobile computing, electronic services, virtual infrastructure, etc. The proposed model is the result of both, theoretical and practical work in the last four years. It defines a general framework for establishment and continuous improvement of e-BCM in insurance companies and includes methods for defining the key components of a business continuity management system: business impact analysis, business continuity risk assessment, and business continuity plan. The results may have broad practical application in insurance companies around the world that uses modern e-business technologies.

Abstract: Past decade revealed testbed based evaluation approaches in networking research as highly important and primarily due to the emphasis on practicality. At the beginning, testbeds were built for a specific project or studying a specific technology. In the context of New Internet, networking and computational testbeds appear of crucial importance for conducting testing of computational tools and new technologies, describing experiments, new platforms and environments. This paper addresses networking testbed enhancements in terms of orchestration, control and virtualization capabilities. It reports about real world evaluation of federated testbed infrastructure in terms of development of ontology-based automatic code generator for experiments. The main idea is to enable experimenters to define semantic description of the experiment through a high-level specification and generate software code that is directly deployable on the testbed federation. As a prerequisite for semantic driven automatic code generation, the ontology framework that gives semantic description of particular experiment domain is developed and its design is discussed in the paper.

Abstract: The reduction of energy losses, particularly non-technical losses, in the distribution network has become one of the priority business goals in companies engaged in distribution of electricity. Accordingly, Electric Power Industry of Serbia use distribution network losses as one of the key elements that indicate the degree of quality of performing its business activities. In this paper, we propose a system for automation of electricity theft detection and monitoring. The proposed system uses analysis of existing data already owned by electric distribution company and does not require expensive equipment. Common model, based on elements of Common Interchange Model, is implemented in order to provide means for the exchange of information between different information systems within Electric Power Industry of Serbia. Proposed systems is already implemented and used in production in different parts of Electric Power Industry of Serbia.

Abstract: Large scale and complex systems that require significant hardware resources are typically designed to utilize distributed computing. This paper presents an architecture of a lightweight multi-agent middleware, aimed to simplify distributed computing operations. This middleware is implemented in JavaScript programming language, specifically in the Node.js framework. It is designed to support simple load balancing, avoid a single point of failure and enable execution of computationally intensive tasks. Furthermore, by supporting standardized agent communication, our agents can transparently interact with agents in existing, third-party multi-agent solutions.

Abstract: In this paper is analyzed the efficiency of image watermarking, based on applying SVD decomposition and its resistance to the presence of impulsive noise. The first part of the paper describes SVD decomposition and SVDU algorithm for watermarking and MDB algorithm for detection and removal of impulse noise. The effect of level inserting of the watermark on visual characteristics of the image with watermark was analyzed in the second part of the paper. Using the data shown in tables and graphs as well as pictures before and after watermarking indicate to the large level of visual image degradation.

Abstract: Clinical decision support systems (CDSS) are designed to assist physicians and other health professionals with numerous clinical tasks, such as establishing a diagnosis or determining the appropriate course of therapy. While much research has been conducted in the field, with various degrees of success, little emphasis has been placed on unified knowledge representation with uncertainty and learning capabilities of a diagnostic system. Non-axiomatic logic (NAL), an Artificial General Intelligence project designed to realize general-purpose logic, provides consistent format for representing and reasoning-learning with knowledge that has diverse degrees of uncertainty. This paper reviews the methods used in design of CDSS, and proposes a CDSS framework based on NAL.

Abstract: Selfdynamisable internal fixator (SIF) is a type of medical device used in internal fixation of long bones. Occasional SIF failure requires an extra surgery, which causes an additional trauma for the patient. To minimize the risk of failure, a methodology was established for optimization of SIF structure and position. It is based on sensitivity studies or structural optimization procedures, performed using finite element method (FEM). Automatic creation of FEM models, for any combination of values of dimensional and positional parameters, is enabled trough creation of a robust and flexible CAD model and establishment of bi-directional associativity between CAD and FEM models. This paper focuses on the results of sensitivity study, performed using the defined methodology.

Abstract: Conjugate to Speech Science are Physiology, Anatomy, Linguistics, Physics, Computer Science, Psychology and so on, that provide tools and methodologies which can decipher the key elements of phonological activity in depth. Additionally, sonic perception provides in synaesthetic terms an acoustic augmentation due to the complex activity of tuned voicing with articulation-related physiological and neuronal dipoles. All these features affect the oral and aural communication channel and enhance the understanding of human communication in terms of Disordered Voicing, Vocal Affections, Music Experience and Human Attitudes.

Abstract: Due to its position and anatomy, human mandible is sensitive to trauma. Fractures of the mandible are different by type, origin (e.g. injuries, tumors), localization (eg. collum, the ramus, angle of mandible, the body of the mandible and alveolar part). Various types of implants are used for the fixation of human bones fractures. Anatomically correct and geometrically accurate personalized implants are necessary in order to improve the quality and duration of the intervention, and postoperative recovery of patients. The aim of this research is to create a geometrical model of the personalized implant plate type which geometry and topology fully matches to the shape of the bone of the patient. The side of the implant, which is in contact with a periosteum outer layer of the mandible, is aligned with the shape of the mandible’s outer surface near the fracture. The obtained model can be used for production of plate implants, and/or for simulation of orthodontist interventions.

Abstract: The dynamic nature and structural heterogeneity of proteins are essential for their functions in live organism. More and more researchers are paying attention to learn the structure of proteins, so obtaining the protein image has become a meaningful problem. During cryo-EM image acquisition, the radiation dose is limited due to several reasons. The electron-beam-induced radiation damage is one of the most important causes of artifacts, sometimes the misinterpretation of micrographs. In the view of above reasons, it is difficult to obtain the complete image through physical methods. Aiming at the problem above, we proposed a feature learning based contextual spectrum inpainting method, which was used to inpaint the missing region of an image through inpainting the image spectrum. Based on the common Deep Convolutional Generative Adversarial Net and simplified Context Encoders, we customized an encoding-decoding generator to minimize the deviation of the generated image spectrum with regard to its latent ground truth. The encoder hierarchically learned the contextual visual features, and the decoder recovered the missing parts of the spectrum with the recombination of the chosen feature maps. The L2 loss function calibrated the error generated by the decoder between original images and recovered images. The recoveries of corrupt image spectrum was presented as an evaluation standard to demonstrate the superiority of our method over some widely used inpaining strategies, and the experiment on the face images from Olivetti Face database [1][2] were performed to validate the generalization of our methods for the face image completion.

Abstract: This paper presents a methodology for evaluation of driving performance based on speeding, acceleration, lane control and safety distance. All these variables are measured in a motion-based driving simulator. We report on a user study in which we obtained the proposed variables for 29 drivers. These results will enable us to propose a general evaluation score of driving performance, which can be used for profiling driver behavior.

Abstract: This study explored the relationship between human biometric signals, such as heart rate, galvanic skin response (GSR) and skin temperature, and cognitive load induced by a secondary cognitive task in a simulated driving environment. NERVteh motion-based driving simulator [1] was used to simulate an immersive driving environment. Microsoft wrist band was used for collection of driver biometric data. Cognitive load was induced with the Delayed Digit Recall Task (n – back task) [2] and the Detection Response task (DRT) [3] was used as a reference measurement. The results show that it is possible to reliably detect changes in cognitive load with a low cost device, such as the Microsoft wristband 2 [4]; however, specific cognitive difficulty levels cannot be differentiated. Galvanic skin response and skin temperature showed to be better indicators for increased cognitive load compared to mean heart rate data, when performing measurements with a low cost wristband.

Abstract: Smart equipment can support feedback in sports training, rehabilitation, and motor learning process. Smart equipment with integrated sensors can be used as a standalone system or complemented with body-attached wearable sensors. The first part of our work focuses on realtime biofeedback system design, particularly on application specific sensor selection. The main goal of our research is to prepare the technical conditions to prove the efficiency and benefits of the real-time biofeedback when used in selected motion-learning processes. The tests performed on two prototypes, smart golf club and smart ski, proves the appropriate sensor selection and the feasibility of implementation of the real-time biofeedback concept in golf and skiing practice. We are confident that the concept can be expanded for use in other sports and rehabilitation.

Abstract: The spatial density of wireless sensor nodes, wearables, and other mobile devices are showing the explosive growth. Without finding new solutions or concepts the existing wireless sensor networks will become a bottleneck. Healthcare and sport wearable sensors come in great varieties regarding their power, data rate, required quality of service, and intended use. Another issue is the number of nodes communicating in one domain. We list the most used wireless technologies with their properties. We present different biofeedback application scenarios in healthcare and sports and match them to the most appropriate existing wireless technology that is expected to sustain scalability in number of nodes or increased data rates for the expected application lifetime.

Abstract: Recent studies indicate that every outdoor physical activity has positive effect on public health and quality of life of every individual [6]. The growth in wireless subscriptions, which has reached over 6 billion wireless subscribers in the world, shows that the use of Information and Technologies (ICTs), more precisely mobile applications intended for improving health and well being (mobile health applications-M-health apps) has become an integral part of everyday urban city life [16]. The aim of this research is to establish the connection between the use of emerging Mhealth apps and outdoor physical activity, public health and well-being of citizens. Important research questiones that guided this study are: Does the use of M-health apps have a positive influence on increasing outdoor sport and recreation activities; How could open public spaces be redesigned in order for people to use them more frequently for recreational activities; Which are the possible ways for upgrading the M-health apps to be more user friendly. The methodology is based on extensive literature review and critical analysis of existing theoretical researches. Quantitative as well as qualitative data presented in this paper is obtained by using the methods of interviews and questionnaire conducted with 300 selected stakeholders from Belgrade. The results of the research are showing general level of physical activity, recreation preferences and habits of citizens in Belgrade, with special focus on the manner, degree and frequency of use of both M-health applications and outdoor spaces. Special contribution of this paper is manifested in presenting the positive effects of Mhealth apps on increasing outdoor physical activity of users in Belgrade, but also in providing specific gueidilines regarding improvement of M-health apps and open public space design according to users’ suggestions.

Abstract: In this paper we discuss the communication standards for Internet of Things applications, with focus on Smart City deployment. Different communication technologies can be used for applications based on the types of devices in the smart environment and their available resources and limitations. A selection criteria for appropriate communication standard also depends on the specific smart environment application. Promising technologies for both indoor and outdoor smart environments are considered. Communication standards comparison is made by considering the various parameters such as throughput, operating frequency bands, nominal range and energy efficiency. Additionally, it was analyzed which standard is the most suitable for appropriate Smart City application.