The fourth industrial revolution, known as Industry 4.0, has tendency to push the boundaries of science and technology. This is especially true for the manufacturing industry. One of the biggest challenges facing the manufacturing industry today is how to make intelligent systems for production with "self-aware", "self-predict and "self-maintain" abilities. Predictive manufacturing systems (PMS) are new intelligent systems that provide these abilities in the production, processes and machines. The PMS combines different technologies and techniques: statistics, data mining, modelling and artificial intelligence methods. These technologies and techniques are used to convert data into information and make predictions about the observed system. This paper provides an overview of the various challenges, existing solutions and benefits of PMS, with a focus on success factors in Industry 4.0.

Cartographic heritage of historical cadastral maps represent remarkable geospatial data. Historical cadastral maps are generally regarded as an essential part of the land management infrastructure (buildings, streets, canals, bridges, etc.). Today these cadastral maps are still in use in a digital raster form (scanned maps). Digitization of cadastral maps is time consuming and it is a challenge for scientists and engineers to find ways to automatically convert raster into vector maps. The process of map digitization typically involves several stages: preprocessing, visual object detection and classification, vector representation postprocessing and extracting information from text. Although neural networks have had a long history of use in the domain, their applications remain limited to extracting the information from text. Recent convergence of advancements in the domains of training deep neural networks (DNN) and GPU hardware allowed DNNs to achieve state-of-the-art results in computer vision applications, beyond hand-written text recognition. This paper provides an overview of different approaches to historical cadastral maps digitization, focusing of the challenges and the potential of using deep neural networks in map digitization.

B. Nikolic, G. Dordevic, D. Milic, N. Milosevic. Performance of SC Receiver over Generalized K Fading Channel in the Presence of Imperfect Reference Signal Recovery // Electronics and Electrical Engineering. -Kaunas: Technologija, 2011. -No. 9(115). -P. 41-46.This paper considers a partially coherent detection of quadrature phase-shift keying (QPSK) signals in a composite generalized K (K G ) fading channel. At the reception the selection combining is applied, while the branches of the combiner are not identically distributed. The extraction of the reference carrier from non-modulated pilot signal is performed in a phase-locked loop (PLL) circuit. The difference between received signal phase and extracted reference signal phase is a stochastic variable with Tikhonov probability density function. The influence of the fading parameters, the standard deviation of phase error and the number of diversity branches on the system performance is examined. The bit-error rate is used as a measure of the reception quality. Ill. 6, bibl. 17 (in English; abstracts in English and Lithuanian). B. Nikolic, G. Dordevic, D. Milic, N. Miloševic. SC imtuvo našumo tyrimas atsižvelgiant į apibendrinantįjį K slopinimą kanale // Elektronika ir elektrotechnika. -Kaunas: Technologija, 2011. -Nr. 9(115). -P. 41-46.Nagrinėjama dalinė koherentinė signalo, moduliuoto kvadratūrine fazine manipuliacija, detekcija atsižvelgiant į apibendrinantįjį K (K G ) slopinimą kanale. Priėmus signalą analizuojamas nešantysis signalas. Nemoduliuoto signalo šaltinis atskiriamas fazinėje kilpoje. Stochastinis kintamasis su Tikhonovo tikimybės tankio funkcija yra pagrindinis skirtumas tarp demoduliuoto signalo ir atskirto signalo šaltinio. Ištirta slopinimo parametrų, standartinio nuokrypio fazinė klaidos įtaka. Il. 6, bibl. 17 (anglų kalba; santraukos anglų ir lietuvių k.).

In this paper an approach to efficient computation of radiation pattern in FDTD simulation environment is presented. A necessary large distance from the radiating object is achieved by multigrid space discretization with unilaterally connected subdomains. A numerical dispersion is reduced using more general complex-envelope finite difference time domain (CE-FDTD) formulation and high order accuracy FDTD schemes where possible. In order to examine how much the introduced algorithm complexity and increased demands concerning computational power and memory are justified by the gain in accuracy, several different scenarios were considered.

No abstract