Dimensional Role Analysis: The Role of Humans and Automation for Increasingly Autonomous Aviation Systems

In this paper, we discuss an analysis method to describe the human and automation roles in increasingly autonomous aviation systems. Specifically, we discuss the relationship between humans and automation in the performance of individual functions in two dimensions, related to their responsibility and authority for the control loops associated with an individual intended function. This work builds upon concepts and terminology proposed by ASTM International in Autonomy Design and Operations in Aviation: Terminology and Requirements Framework (TR1) published in 2019 and is based in part on the on-going work of an ASTM Working Group (WK76044) under F39 Aircraft Systems. TR1 presented a Contextual Framework for analysis of automated aviation functions, and the working group is developing a practice for its implementation.The role of the human vs. the role of automation, which has been extensively discussed in the past, is fundamental to the implementation of the Contextual Framework. As such, it is worth a fresh look, particularly against the backdrop of the Contextual Framework. Our work has led to an analysis of the system state, which at any given time is defined by the role of the human and the role of automation in a two-dimensional space. We refer to this as Dimensional Role Analysis which uses a directed graph approach to clearly depict the nominal operating mode, the revisionary modes, and the triggering events that would cause transition. The analysis approach helps to establish clear delineations as to the responsibility and authority of humans and automation in the performance of individual intended functions. This work makes a clear distinction between authority, responsibility, and accountability consistent with the work of ASTM.Using the concepts and language presented in this paper could add clarity to discussions in the aviation community, as proponents seek to certify and obtain operational approval for increasingly autonomous systems in aviation. While this paper focuses on aircraft control, these approaches could also apply to other safety critical systems.