Abstract. In this paper, we show an approach to build deep learning algorithms for recognizing signals in distributed fiber optic monitoring and security systems for long perimeters. Synthesizing such detection algorithms poses a non-trivial research and development challenge, because these systems face stringent error (type I and II) requirements and operate in difficult signal-jamming environments, with intensive signal-like jamming and a variety of changing possible signal portraits of possible recognized events. To address these issues, we have developed a two-level event detection architecture, where the primary classifier is based on an ensemble of deep convolutional networks, can recognize 7 classes of signals and receives time-space data frames as input. Using real-life data, we have shown that the applied methods result in efficient and robust multiclass detection algorithms that have a high degree of adaptability.

Abstract.A new method of symbolic analysis based on finite discretization of velocity-curvature space is proposed. A minimum alphabet is introduced in a natural way, and a number of initial analytic measures are defined that make it possible to study the structure of discrete mapping dynamics. The proposed method is tested by application to a system of two unidirectionally coupled logistic maps. It is shown that this method can be used to reveal and study changes in the structure of attractors. In the given example, features in the attractor structure of the driven subsystem are studied upon its escape from the identical synchronization regime.

No abstract

A new approach to analysis of the synchronization of chaotic oscillations in two (or more) coupled oscillators is described that makes it possible to reveal changes in the structure of attractors and detect the appearance of intermittency. The proposed method is based on a symbolic analysis developed previously in the velocity-curvature space of multidimensional sequences and maps. The method is tested by application to a Lorentz system. The results confirm the informativity of the analyzer and reveal specific features of changes in the structure of an attractor of the three-component test system.Synchronization of chaotic oscillations [1,2], which is among the most fundamental concepts of the theory of nonlinear dynamics and chaos, can take place via several routes, including the complete (identical) [3], frequency [4], phase [5], generalized [6], lag [7], timescale [8], and antiphase type [9]. At present, investigations are aimed at (i) considering various types of synchronization from a common standpoint and (ii) seeking for new types of synchronous behavior.Previously, the author proposed [10] a new method of symbolic analysis that was based on finite discretization of the velocity-curvature space and a minimum alphabet introduced in a natural way. According to this, the sequence

Abstract.A new approach is proposed to the integrated analysis of the time structure of synchronization of multidimensional chaotic systems. The method allows one to diagnose and quantitatively evaluate the intermittency characteristics during synchronization of chaotic oscillations in the T-synchronization mode. A system of two identical logistic mappings with unidirectional coupling that operate in the developed chaos regime is analyzed. It is shown that the widely used approach, in which only synchronization patterns are subjected to analysis while desynchronization areas are considered as a background signal and removed from analysis, should be considered as methodologically incomplete.