Vol 4, No 3 (2024)

The paper discusses broadband linear discrete antennas focused in the near radiated field zone. It is shown that when taking into account the properties of focused electromagnetic fields, it is necessary to take into account the methods of using the energy of the focused electromagnetic field in various operating modes. It is shown that for different modes, the spatial distribution functions of focused electromagnetic fields are different. The main properties of electromagnetic fields focused in the near radiated field zone are proposed and their quantitative estimates are given for various application modes.

A new amplitude fiber-optic electromagnetic field sensor based on a micron-sized cavity filled with 5SB liquid nematic crystals has been proposed and tested experimentally. The sensitivity of the sensor has been determined, and options for further improvement have been suggested.

The developed method of monitoring pipeline (PL) leaks involves the use of a number of fiber-optic accelerometers installed along the PL for continuous quasi-distributed monitoring of its dynamic behavior. To select the design of a fiber-optic accelerometer, the task was set to analyze the existing methods and means of monitoring vibration caused by leaks, based on the use of sensitive elements of the fiber Bragg grating (FBG) type, as the most promising: resistant to electromagnetic interference and having optomechanics operating in a wide range of temperatures. Based on the analysis results, it is possible to make a decision on the influence of the sensitive element on the metrological and technical and economic characteristics of the accelerometer as a whole, and in particular, on the characteristics determining its design. The general formulation of the analysis problem can be divided into three main parts: the possibility of leak detection (sensitivity), spatial characteristics of the analysis - resistance to transverse and arbitrary effects different from axial ones, and, finally, the implementation of the accelerometer operating point control function. Based on the analysis results, a uniaxial accelerometer with two parallel fibers was selected according to the specified criteria, into which one FBG is built in before and after the inertial mass. During the analysis, attention was paid to the transition from optoelectronic methods of accelerometer interrogation to microwave photonic ones. The possibility of such a solution is confirmed by the use in the accelerometer of a new type of sensitive element instead of FBG - an addressable fiber Bragg structure (AFBS) of wave - or phase π with one address, as well as their multi-address versions Nl and Nπ. To interrogate the AFBS, an addressable microwave photonic interrogator should be built, which is significantly cheaper in terms of the mass-dimensional parameters of the size, weight, cost and power consumption, compared to the optoelectronic one, and also better in terms of metrological and technical-economic characteristics. In the development of this type of accelerometers, the paper proposes the use of new AFBS of a combined type - (l+l/π)-AFBS, the main address frequency of which is not clearly expressed, i.e. cannot be determined by measurements only for transmission or only for reflection. This confirms the possibility of constructing lines of quasi-distributed accelerometers or their arrays of complex PL topology, which is difficult to implement when using single- and multi-address AFBS that do not allow combined inclusion.

In this work, we investigated the application of neural networks for handwritten signature recognition, taking into account the dynamics of their input. We considered three types of architectures: Multilayer Perceptrons (MLPs), Long Short-Term Memory networks (LSTMs), and Convolutional Neural Networks (CNNs). We conducted a comparative analysis of the effectiveness of these architectures on the MCYT Signature 100 dataset, which contains signatures from 100 authors and their forgeries. Key hyperparameters, such as the optimizer (RMSProp, Adam, SGD), batch size, and number of layers, were varied for each architecture during the study. The results show that Convolutional Neural Networks (CNNs) achieve the best accuracy, outperforming MLPs and LSTMs. The proposed method has practical value for automating signature verification and can be applied in security systems and electronic document management.

Television, being the most accessible and popular means of mass media in the world, plays a major role in the development of human society. In the context of globalization and information society, interest in the historical origins of a particular phenomenon is natural. In the history of world television, a special place is occupied by the inventions of the first color television system and audio-video magnetic recording, which belong to Russian inventors A.A. Polumordvinov and A.M. Ponyatov, whose fate was connected with Kazan and Kazan University – the oldest university in Russia, which turned 220 years old in 2024.