An autonomous quadrotor avoiding a helicopter in low-altitude flights

摘要

Recent advances in sensing and computing technology has made unmanned aerial vehicles/systems (UAV/UAS) low cost but still increasingly capable of executing complex missions in challenging environments. They have gained popularity in a vast range of civilian applications, including search and rescue, disaster relief, and filming. Recently, the Federal Aviation Administration (FAA) has issued the Notice of Proposed Rulemaking (NPRM) on UAS certifications [1], indicating that a large number of UAS will be present in the National Airspace System (NAS) in the near future. The NASA UTM project is an effort on enabling low-altitude UAS flights [2]. Big value envisaged by Amazon Prime Air [3] happens only when the drone is able to fly itself tens of miles from the distribution center to people’s homes autonomously. One prerequisite for such flights is collision avoidance. Our research aims at drones that travel in class G airspace. This article is an exploration of the drone collision avoidance problem in urban areas. The main contribution is a safety control framework that enables a UAS to perform collision avoidance with manned aircraft during autonomous navigation.