SLDS - Self-Learning Drone Surveillance

Important results from the project

The goal of the project was to develop a system for fully autonomous drone detection with the following properties: 1) detect drones within our supported frequency bands, 2) identify different and new drone models, 3) distinguish drones of the same model, and 4) locate them in 3D in a TDOA network. These objective have been achieved through field-tested systems that captures radio signals via TDOA and uses an AI model that automatically recognizes new wireless drone protocols.

Expected long term effects

There is a constant game of cat and mouse between those who want to fly drones undetected and those who monitor the airspace. This arms race has intensified in recent years as both drones and anti-drone systems have become more sophisticated. AI-assisted systems, like the ones we have developed, can give the monitor an advantage by automating the analysis of the spectrum, something that previously required human experts and extensive forensic work.

Approach and implementation

SLDS aimed to achieve self-learning detection of new wireless drone protocols - and we have achieved that. In the field, the system can independently detect a radio protocol previously unknown to the system. The approach was clear and has worked well. Skysense was responsible for a system that registers radio traffic in the lower airspace (= presumed drones). KTH developed the system for machine learning. Securitas tested the system and evaluated the results from an operational perspective.