No abstract

No abstract

The development of a new adaptive system of radar data semantic analysis with their non-stationarity, which is based on both numerical and logical methods of multiscanning processing of signals and methods of artificial intelligence using fuzzy transformations of the universe of signals and signal images, is proposed. The possibility of its hardware and software implementation is considered. The results of computer modeling, theoretical and experimental researches with processing of real radar signals are presented. The elements of logical analysis and algebra of finite predicates (AFP) are selected as mathematical apparatus. As experimental studies show, AFP is an appropriate tool for logical-mathematical constructions, with which it's possible to describe the radar operator actions. The basic concepts of Boolean algebra and graph theory are also used. The practical value of the work is: a method for formalizing the processes of perception and transformation of signals and signal images, algorithms and software are intended for information radar systems with natural-language intellectual interface; also for support the design of information structures. Mathematical and software results can be used in the systems of automatic processing of radar information, particularly, in the intelligent radar and radio-electronic systems and complexes for monitoring of mobile air and ground objects.

The paper presents results of development of the method and experimental studies of the system for automatic detection of radar signals of aerial objects and their recognition with the processing of real records in surveillance radars. The relevance of this work consists in creation of algorithms for automatic information processing to ensure effective detection of useful signals due to accumulation of signal (energy) and semantic information. The method is based on the definition of semantic components at the stage of formation and analysis of the symbolic model of signals from point and extended air objects. Signal information is described by the predicate function of process knowledge of the formation and analysis of a symbolic model of a burst of impulse signals from point-like mobile aircraft such as an airplane, a helicopter, a UAV, and from extended atmospheric formations such as angel-echoes, clouds. As a result of semantic analysis of symbolic images of signal marks, classification distinctive features of air objects were obtained. The semantic components of the decision-making algorithm, similar to the decision-making algorithms used by the operator, have been investigated. In the developed algorithm, signal information is described by a predicate function on the set of signal mark pulse amplitudes that have exceeded a certain threshold value. Recognizing of aerial objects is carried out by solving the developed equations of predicate operations. The verification of the developed method was carried out on real data obtained on a survey radar of the centimeter range (pulse duration was 1 μs, probing frequency wass 365 Hz, survey period was 10 s). Based on these data, the types of characteristic marks of radar signals are modeled. According to the results of the experiments, they were all correctly identified.