T HE DATA CHALLE NGE
A recent NCI Agency development challenge set participants the task of improving the detection of Unmanned Aircraft Systems. Mike Bryant asks Adelica Ndoni, a junior data scientist at the NCI Agency, to explain what the judges were looking for 70
A particular highlight of the recent International Conference on Military Communications and Information Systems (ICMCIS), organized by the NCI Agency in collaboration with NATO’s Science and Technology Organization and held on 4-5 May, was a special session in which the winners of the Agency’s latest challenge presented their solutions to a thorny problem. This year, the participants had to focus on the application of artificial intelligence (AI) and machine learning (ML) in ways that have the potential to enable commanders to make better decisions faster, by providing more comprehensive levels of situational awareness. Participants in the Class I Unmanned Aircraft Systems (UAS) Challenge were tasked with offering their solution to tracking, classifying and identifying Class I UAS flying within a protected zone using the sensor data made available. Class I UAS weigh less than 150kg; the classification covers most hobby drones, but that is not to say they cannot present a significant threat when flown in prohibited airspace.
Those who took up the challenge could fuse together several sources of data provided by the Agency to identify and track the drones. These sensors included radar and radio direction-finding, data from which had to be assessed to confirm the presence of a Class I UAS – as opposed to, for example, a bird or any other unidentified object flying in the test zone – then classify and identify it (based on a number of features typically found in UAS). If classified as a drone, the UAS had to be tracked and its ongoing speeds and locations recorded. It then had to be identified as a particular type of drone: for instance, a DJI Mavic Pro, DJI Phantom 4 Pro etc. The UAS Challenge formed part of a larger research and development effort by the NCI Agency, aimed at developing effective remote sensing technologies that are suitable for detecting, tracking and identifying Class I UAS. The challenge supports the efforts that NATO and, hence, the NCI Agency is making to protect its people, facilities and missions against