M. E. Ivanov1

1 Saint-Petersburg State University of Aerospace Instrumentation (St. Petersburg, Russia)

1 sindbad1995@list.ru

The effectiveness of modern automatic control systems for the movement of aircraft depends on the ability to solve flight tasks of any complexity class with specified criteria of accuracy and speed. At the same time, the results of the measurement systems are affected by the level of a priori uncertainty, which inevitably arises due to insufficient completeness of flight information. To reduce the degree of a priori uncertainty, the transmission of information by means of radio signals from various sources is widely used in solving specific tasks: for example, signals received from orbiting satellites make it possible to correct the readings of the onboard inertial navigation system. The consequence of the large number of radio signals received by aircraft sensors is the task of recognizing their types for further processing and use, which is proposed to be solved in the framework of the study using one of the methods of artificial intelligence – an artificial neural network.

In this article, the mathematical apparatus of artificial neural networks, one of the methods of artificial intelligence is used: structural and functional synthesis, training for calculating the weight coefficients of synapses (the method of training is the reverse propagation of error). The mathematical model of a neural network is a multilayer Rosenblatt’s perceptron with a threshold activation function.

The aim of the article is to establish the possibility of solving the problem of recognizing radio signals using an artificial neural network such as a single-layer Rosenblatt perceptron with a threshold activation function, to demonstrate the methodology for the synthesis of such a neural network, as well as to develop recommendations for improving the level of its functionality.

The structural-functional and parametric synthesis of the target artificial neural network has been carried out for recognition of both deterministic signals (from orbital satellites, from control stations and automatic direction finders) and stochastic noises. The training of a synthesized neural network has been implemented by the error back propagation method during a finite number of iterations.

The results of the work can be used in the construction of multi-level intelligent automatic control systems capable of solving flight tasks of varying degrees of complexity: measurement of flight parameters, landing on the runway in the automatic mode, collision warning in the air and with the ground, adaptive adjustment and correction of measured parameters.

Ivanov M.E. Application of an artificial neural network for recognition of radio signals by aircraft receivers. Information-measuring and Control Systems. 2023. V. 21. № 4. P. 19−24. DOI: https://doi.org/10.18127/j20700814-202304-03 (in Russian)

1. Andrejchikov A.V., Andrejchikova O.N. Intellektual'nye informatsionnye sistemy i metody iskusstvennogo intellekta. M.: NITs INFRA–M. 2021. (in Russian)
2. Skakovskaya A.N., Moiseev D.V., Anisimova M.V. Ispol'zovanie iskusstvennykh nejronnykh setej dlya raspoznavaniya obrazov. Sb. materialov 26-j Mezhdunar. Krymskoj konf. «SVCh-tekhnika i telekommunikatsionnye tekhnologii». Sevastopol'. 2016. T. 4. S. 825–830. (in Russian)
3. Sergienko A.B. Tsifrovaya obrabotka signalov. SPb.: Piter. 2003. (in Russian)
4. Selivanova L.M., Shevtsova E.V. Inertsial'nye navigatsionnye sistemy. Ch. 1. M.: MGTU im. N.E. Baumana. 2012. (in Russian)
5. Rozenblatt F. Printsipy nejrodinamiki: Pertseptrony i teoriya mekhanizmov mozga: Per. s angl. V.Ya. Altaeva i dr. Pod red. S.M. Osovtsa. M.: Mir. 1965. (in Russian)
6. Levashov O.V. Ot Rozenblatta do Khintona: plyusy i minusy sovremennykh nejronnykh setej. Sb. materialov III Mezhdunar. konf. «Nejrokomp'yuternyj interfejs: nauka i praktika». Samara. 2017. S. 65–69. (in Russian)
7. Plotnikov D.V. Vychislitel'nyj intellekt v zadache raspoznavaniya obrazov. Aktual'nye problemy aviatsii i kosmonavtiki. 2017. T. 2. № 13. S. 52–54. (in Russian)