Demonstration of closed-loop airborne sense-and-avoid using machine vision

摘要

The ARCAA DSA program has recently demonstrated full closed-loop aircraft sense-and-avoid using vision-based sensor technology. Real-time onboard detection and tracking was accomplished by harnessing the parallel processing capabilities of GPU-based hardware. ARCAA''s DSA program aims to develop a cost-effective collision avoidance capability to facilitate the routine operation of UASs in the national airspace as well as enhance manned aviation safety by complementing human pilot “see-and-avoid.”