Cherepovets State University specialists are trying to create a system from a group of microphones to determine the distance and direction to the quadcopter by sound. This was reported to socialbites.ca in the press center of the operator of the NTI Platform acceleration programs.
The developed module is a software and hardware complex consisting of a microphone array and artificial intelligence. Microphones monitor the acoustic environment and transmit information about the surrounding sounds to a computer, where it is processed by a neural network algorithm. AI can isolate the sound of a flying drone during analysis and determine the distance and direction to it.
Nikita Efimov, a graduate of the Department of Cyber-Physical Systems, one of the developers of the module, explained to socialbites.ca that drones are currently tracked by passive means (video cameras and thermal imaging devices) and active means (radio direction finders). . However, according to experts, they have their drawbacks.
“Visual analysis systems, for example, are large and active circuits detect themselves on the air, and therefore the observed object can go into radio silence or damage the search engine,” said Yefimov.
According to Efimov, a development-like sound module under discussion cannot be detected because it does not emit signals. It can also work in complete darkness and has a smaller size.
Tests of the system will begin before the end of 2023. Experts will then be able to identify precise performance indicators. Currently, developers expect the microphone array to be able to detect objects within a one kilometer radius with 94 percent accuracy.
before in Russia developed Neural network to facilitate working with Excel spreadsheets.