Introduction: Insufficient physical activity is one of the leading public health problems in the world, but also in Bosnia and Herzegovina. Modern civilization is characterized by a significant decrease in physical activity, and the number of people whose lifestyle can be called sedentary has never been higher, which is especially emphasised among children and adolescents. Aim of the study is to examine public health significance of physical activity on the occurrence and the degree of obesity in children and adolescents in primary and secondary schools and to determine the applicability of the Fels questionnaire on physical activity of children in rural areas of Bosnia and Herzegovina. Methods: We used a transversal research method of a cross-sectional study at a one-time point, and for obtaining results we used the Fels physical activity questionnaire for children and measurement protocol. Results: 276 primary and secondary school students in two cities participated in this survey. Respondents in Busovaca are more physically active than their peers in Sarajevo. One-third of the total number of respondents is overweight and obese, and respondents in Sarajevo are significantly more nourished than their peers in Busovaca. The Fels questionnaire is conditionally applicable, especially in rural areas. Conclusion: This study confirmed that the Fels questionnaire for assessing the level of physical activity for children and young people, which is the general instrument for research of physical activity in children, is too generalized because it is based on a homogeneous urban population.

One of the major goals in the European Union for reducing greenhouse gas emissions is the electrification of heat. Therefore, it is expected that the winter peak demand will rise significantly in the next few years. Demand Response could play an important role in reducing the need for network reinforcements by providing flexibility. The major motivation behind this paper is to evaluate the difference in demand flexibility between temperature-dependent consumers using electricity for heating and consumers using other energy sources. In this paper, temperature-dependent consumers are first identified by analyzing their smart metering data with machine learning. Further, the response of consumers is evaluated using probabilistic baseline models. The results show that heat electrification will increase the demand during low temperatures, whereas these consumers will also be able to offer far more flexibility during low temperatures and high demand. To the best of our knowledge, there is no empirical study, that would investigate these using state of the art methods in such detail. The paper presents part of the analyses that were carried out after the real demand response program in the scope of the Slovenian-Japanese NEDO project.

This paper investigates a secure data exchange between many small distributed consumers/prosumers and the aggregator in the process of energy balancing. It addresses the challenges of ensuring data exchange in a simple, scalable, and affordable way. The communication platform for data exchange is using Ethereum Blockchain technology. It provides a distributed ledger database across a distributed network, supports simple connectivity for new stakeholders, and enables many small entities to contribute with their flexible energy to the system balancing. The architecture of a simulation/emulation environment provides a direct connection of a relational database to the Ethereum network, thus enabling dynamic data management. In addition, it extends security of the environment with security mechanisms of relational databases. Proof-of-concept setup with the simulation of system balancing processes, confirms the suitability of the solution for secure data exchange in the market, operation, and measurement area. For the most intensive and space-consuming measurement data exchange, we have investigated data aggregation to ensure performance optimisation of required computation and space usage.

Electricity sector has been facing many changes over the last two decades due to rise in penetration of distributed energy resources that significantly affect the operations of distribution grids. Increase in intermittent electricity production from renewable energy sources, requires activating the flexibility contained in the distributed energy resources. Local electricity market as well as demand response present a mechanism to utilize this flexibility. In this paper, we analyze potentials of energy exchange within energy community created at medium voltage feeder of Elektroprivreda BH – d.d. Sarajevo. We use software tool PVSOL Premium to model prosumers and Python for the analysis of power flows and voltage conditions. As a result, we propose the energy community interaction matrix providing the information about prosumers and consumers as a foundation for automation of local energy exchange within the energy community.

Deep scattering spectrum consists of a cascade of wavelet transforms and modulus non-linearity. It generates features of different orders, with the first order coefficients approximately equal to the Mel-frequency cepstrum, and higher order coefficients recovering information lost at lower levels. We investigate the effect of including the information recovered by higher order coefficients on the robustness of speech recognition. To that end, we also propose a modification to the original scattering transform tailored for noisy speech. In particular, instead of the modulus non-linearity we opt to work with power coefficients and, therefore, use the squared modulus non-linearity. We quantify the robustness of scattering features using the word error rates of acoustic models trained on clean speech and evaluated using sets of utterances corrupted with different noise types. Our empirical results show that the second order scattering power spectrum coefficients capture invariants relevant for noise robustness and that this additional information improves generalization to unseen noise conditions (almost 20% relative error reduction on AURORA 4). This finding can have important consequences on speech recognition systems that typically discard the second order information and keep only the first order features (known for emulating MFCC and FBANK values) when representing speech.

Due to limited computational resources, acoustic models of early automatic speech recognition ( ASR ) systems were built in low-dimensional feature spaces that incur considerable information loss at the outset of the process. Several comparative studies of automatic and human speech recognition suggest that this information loss can adversely affect the robustness of ASR systems. To mitigate that and allow for learning of robust models, we propose a deep 2 D convolutional network in the waveform domain. The first layer of the network decomposes waveforms into frequency sub-bands, thereby representing them in a structured high-dimensional space. This is achieved by means of a parametric convolutional block defined via cosine modulations of compactly supported windows. The next layer embeds the wave-form in an even higher-dimensional space of high-resolution spectro-temporal patterns, implemented via a 2 D convolutional block. This is followed by a gradual compression phase that selects most relevant spectro-temporal patterns using wide-pass 2 D filtering. Our results show that the approach significantly out-performs alternative waveform-based models on both noisy and spontaneous conversational speech ( 24% and 11% relative error reduction, respectively). Moreover, this study provides empirical evidence that learning directly from the waveform domain could be more effective than learning using hand-crafted features.