Key points are not available for this paper at this time.
While academia and industry have shown that they can develop systems that exhibit high levels of autonomy across multiple domains (air, sea, and land systems), as of now, regulations do not exist to allow a system to operate autonomously (without a human in or on the loop to ultimately be accountable for the actions of the system). Without these regulations, a truly autonomous system will remain a demon-stration program or simply science fiction. The United States Navy has publicly announced that it intends to dramatically increase its fleet of uncrewed aerial systems. To support an ongoing effort for fielding autonomy, researchers have been trying to define the problem and identify possible solutions. This survey paper details the work that has gone into generating certification evidence for artificial intelligence/machine learning algorithms to replace humans within naval aviation. It is hoped that the lessons learned and standards developed can be used to spread autonomous functionality to other domains.
Costello et al. (Tue,) studied this question.