For the last two decades, artificial neural networks (ANNs) of the third generation, also known as spiking neural networks (SNN), have remained a subject of interest for researchers. A significant difficulty for the practical application of SNNs is their poor suitability for von Neumann computer architecture, so many researchers are currently focusing on the development of alternative hardware. Nevertheless, today several experimental libraries implementing SNNs for conventional computers are available. In this paper, using the RCNet library, we compare the performance of reservoir computing architectures based on artificial and spiking neural networks. We explicitly show that, despite the higher execution time, SNNs can demonstrate outstanding classification accuracy in the case of complicated datasets, such as data from industrial sensors used for the fault detection of bearings and gears. For one of the test problems, namely, ball bearing diagnosis using an accelerometer, the accuracy of the classification using reservoir SNN almost reached 100%, while the reservoir ANN was able to achieve recognition accuracy up to only 61%. The results of the study clearly demonstrate the superiority and benefits of SNN classificators.

Sensors based on chaotic oscillators have a simple design, combined with high sensitivity and energy efficiency. Among many developed schemes of such sensors, the promising one is based on the Duffing oscillator, which possesses a remarkable property of demonstrating chaotic oscillations only in the presence of a weak sine wave at the input. The main goal of this research was to evaluate the maximal sensitivity of a practically implemented metal detector based on the Duffing oscillator and compare its sensitivity with conventional sensors. To achieve high efficiency of the Duffing-based design, we proposed an algorithm which performs a bifurcation analysis of any chaotic system, classifies the oscillation modes and determines the system sensitivity to a change in different parameters. We apply the developed algorithm to improve the sensitivity of the electronic circuit implementing the Duffing oscillator, serving as a key part of a three-coil metal detector. We show that the developed design allows detecting the presence of metal objects near the coils more reliably than the conventional signal analysis techniques, and the developed detector is capable of sensing a large metal plate at distances up to 2.8 of the coil diameter, which can be considered a state-of-the-art result.

Maintaining ecosystems is one of the current leading public concerns. Some of the pressing problems that threaten water resources are pollution of water bodies with floating debris, illegal extraction of water resources, wear and tear of underwater communications. Therefore, the creation of special technical solutions is urgent. This paper reports a model-based design of an unmanned surface vehicle (USV), purposed to control and maintain the oxygen level and parameters such as acidity and the water temperature in rivers, lakes, inland waterways, and coastal waters. The developed USV navigation autopilot is described as a system with two inputs and one output. The autopilot selected is an adaptive controller based on the concept of proportional, integral, and derivative (PID). The autopilot is implemented on the STM32 microcontroller and allows precisely maintaining a given course, adjusting the speed and angle of rotation during wind drift and other influences. The new technique for sensor calibration and data acquisition is described. Simulation results are given, showing the performance of the autopilot algorithm in response to variations in the environment. The numerical experiments of the model have presented the result of confirming the sufficient correspondence between prototype characteristics and simulation results. Finally, thorough field trials were performed to verify the reliability and precision of the proposed solutions. The developed unmanned surface vehicle can be used for environmental monitoring (water sampling, hydrography, patrolling water areas). In turn, the solutions obtained will be suitable for the design of other USV of different sizes.

The purpose of this work is to predict the destruction of reinforced concrete beams that are subjected to three- point physical loads. A decrease in the number of acoustic emission signals with increasing load was taken as a criterion for the onset of destruction. Polynomial regression is considered as a forecasting method. Using the Sklearn Python library, the coefficients of the polynomial are fitted. The degree of the polynomial is found empirically. The developed method has shown good results in two practical experiments.