Drone Detection Using Sparse Lidar Measurements

Unmanned aerial vehicles (UAV) have been serving people, fulfilling various roles in research, industry and military, not to exclude personal entertainment. They have also been in use more often for malicious purposes, such as annoying neighbors, disrupting air-travel, attacking people or infrastructure. Therefore, motivated particularly by malicious uses, the research community has been interested in proposing solutions for detection of UAVs. In this work, we present a ground based aerial target detection system using lidars, relying on sparse detections rather than dense point clouds, with the aim of not only detecting but also estimating their motion and active tracking. We lay a theoretical groundwork to analyse the performance of such a lidar based detection system and help understanding its limitations. The work is complemented with field experiments utilizing a state-of-the-art lidar and several drones of different sizes, showing effectiveness in both detection and tracking using sparse measurements.