В Москве создали навигационный комплекс для беспилотников

The development will allow determining the location of the drone in case of signal loss

A navigator for unmanned aerial vehicles (UAVs) based on artificial intelligence (AI) has been developed at the Moscow Aviation Institute (MAI). It will enable drones to determine their location in case of satellite communication loss. This was reported by the university's press service.

The research stand for new navigation systems was created on the platform of a passenger car. Specialists moved around in it in dense urban environments and collected data to train the AI. The neural network was trained not only to recognize errors, but also to identify unreliable satellite signals for their exclusion from future navigation solutions.

It was found that the new device for UAV navigation is more effective than the Kalman filter (a mathematical algorithm that allows estimating the state of a system based on incomplete, noisy information).

This is because the conditions under which this mathematical algorithm works are violated here. Artificial intelligence does not have such a strict mathematical model as an optimal filter, where predictions are made on the basis of dozens of mathematical equations, but it has experience, and repeated experience, gained from dozens or even hundreds of passes, of how the navigation system behaved, and based on its previous experience, it can give an accurate forecast of how the navigation system will behave.
Konstantin Veremeenko, Head of the Research Department of Institute No. 3 of MAI

Earlier www1.ru reported that in Sevastopol they will begin to create software for autonomous drone flights.

Read materials on the topic:

Will help in the search for people and the protection of territories: a software and hardware complex for drones has been created in Russia

"Smart" UAV will hit the enemy: the latest FPV drone "Veter" with artificial intelligence was used at the front

AI will put enemy drones into a "Stupor": Russia has new complexes to protect objects from UAVs

Now on home