Machine Learning and Radar Detect Drones

A mockup of what a radar antenna detecting a drone in an urban setting might look like. (Duke University)

If radar signals move down from the clouds and into a city’s streets, there are suddenly many objects that can be mistaken for one another. Drones can easily be hidden on radar displays among slowly moving cars, bicyclists, a person jogging, or even the spinning blades of an air conditioning unit.

Radar, machine learning, and specialized hardware were combined to make a drone surveillance system with sufficient range to allow drones to be detected and stopped before they reach a protected area in a city.

The result is a system that not only spots drones with 98% accuracy but a system that also consumes 100 times less energy than a similar GPU-based system would, all while maintaining the performance and speed required to work in real time.

Source