Roselectronics has developed a neural network for optical means of drone detection

Illustration generated by Stable Diffusion neural network

Cyclone Scientific Institute of the Roselektronika holding has developed a neural network for optical detection systems for unmanned aerial vehicles (UAVs) in protected areas. The technology makes it possible to increase the range of such systems by 40 per cent, which significantly increases their efficiency, Rostec said.

“A team from Cyclone created a neural network as part of a technological marathon to develop digital solutions for government agencies, commercial organisations and regions, known as Leaders of Digital Transformation.” The participants were tasked with creating an aerial object detector capable of detecting any flying objects and classifying them by threat level. The expert jury highly appreciated the project, which at the end of the competition took the prize,” the report says.

The main purpose of the new neural network is to automate the work of optical detectors that scan the sky near the protected objects. The introduction of computer vision and artificial intelligence methods makes it possible to identify aircraft in a timely manner, determine their type and classify the degree of threat. When a potential threat is detected, the operator is notified to take measures to suppress the UAV. A significant advantage of the Cyclone development is the possibility of fully autonomous operation of drone countermeasures systems, which reduces the need for constant human control.

According to Yuri Koval, technical director of Cyclone, the created neural network demonstrates high efficiency and has great potential for development in the field of security. “According to our rough estimate, the created neural network, compared to similar IT solutions, is capable of increasing the range of UAV detection systems by about 40 per cent,” he said.

In developing the neural network, specialists from Cyclone used a combination of several optimised neural network models. This approach allowed to achieve a significant accumulative effect, increasing the accuracy and range of detection. The technology can be applied to various critical infrastructure facilities, as well as to protect private territories from illegal perimeter breaches.

1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)
Loading...