CLASS U-space drone test flight results for non-cooperative surveillance using an L-band 3-D staring radar

Non-cooperative surveillance of drones is an important consideration in the EU SESAR vision for the provision of U-space services. Aveillant Gamekeeper multiple beam staring radar utilises extended dwell to be able to detect small drones at the range of several kilometres. However, target discrimination is necessary with such surveillance system as the increased detection sensitivity against low RCS targets extenuates the problem of false reports of targets such as birds and surface objects such as vehicles etc. Machine learning classifiers are used to remove confuser targets such as birds to provide real-time tracks of drones. Field trials from SESAR CLASS project live drone flights against several test scenarios for U-space are used to train and test a decision tree classifier working on both trajectory and micro-doppler features. Results show that a high level of classifier accuracy is achieved across a range of flight profiles for a rotary wing drone.