An approach for determining the correlation functions of the noise modulation function under the influence of modulating (or more often multiplicative) noise is described. Both stationary and slow noise is considered as influencing multiplicative noise. Correlation functions of noise modulation and fluctuations of the latter can be found by the four-dimensional characteristic function of phase and amplitude distortions. We have derived expressions for the abovementioned correlation functions affected by static multiplicative noise for the normal law of variation of amplitude and phase distortions in the case of their uncorrelation and cross-correlation. The correlation functions of noise modulation and noise modulation function fluctuations under the influence of slow modulating noise were determined for the case when the phase and distortion amplitude variations are independent. The expressions of the correlation functions of noise modulation and its fluctuations for correlated and normal distribution of phase and amplitude distortions under the influence of slow modulating noise are obtained.
The operation of anomaly detection systems in modern computer networks, as a rule, is associated with the processing of large amounts of traffic. With the increase in the scale of computer networks and the growing complexity of network attacks, it becomes necessary to detect multi-stage attacks in real time. This paper presents the architecture of a multi-stage anomaly detection system. The features of the system are the use of generative adversarial neural networks and the minimization of processed traffic using an attacker’s behavior model. The described architecture has a multilevel structure and allows monitoring in distributed computer networks.
The study presents an analysis of universal JFET and CMOS differential stages (DS) circuits. Such cascades provide a smaller range of transistor currents (10-100 $\mu$A) compared to classical cascades. The described DSs are implemented on combined process technologies, and therefore can be designed on JFET, BJT or CMOS chips. In addition, gallium arsenide, diamond gallium nitride, thin-film TFT and silicon carbide transistors. A computer simulation of the static modes of the developed DSs and their current dependences is presented. The results of GaAs DS modeling and its frequency dependence are presented.
This article presents the research results of JDE algorithm for detection and tracking athletes in video data. The developed convolutional neural network was trained and tested on the NVIDIA DGX-1 supercomputer. To analyze the quality of the model, MOTA metric was used, which is related to how the human eyes track targets in the video stream. The detection and tracking people on video data is relevant to many tasks of computer vision, in particular, in sport to collect statistics about players. The quality of JDE algorithm was assessed by the video fragments of basketball games filmed in the sports halls at P.G. Demidov Yaroslavl State University.
The article presents the results of extracting the envelope of radio impulse signals using the digital processing methods. Three methods of extracting the envelope are studied in the work: correlation function; the Hilbert transform; and the new approach in which the combination of correlation function and Hilbert transform are used. For the research we used the measured reflectograms of the multilayer liquid. As a result of processing and comparing a variety of experimental characteristics the accuracy of determination of Hilbertcorrelation combination sequence was better than standard methods of processing. The resulting waveform is easier to analyze and process, but the processing need more time to calculate.
