Frameworks for Supervisory Control: Characterizing Relationships with Uninhabited Vehicles

Abstract

This article reports efforts to devise frameworks for quickly and easily characterizing 13 different supervisory control applications developed and tested by the international participants in a North Atlantic Treaty Organization working group on supervisory control of multiple uninhabited vehicles (UVs). Multiple prior frameworks for supervisory control are reviewed, and two novel ones (a more complex one involving seven descriptive parameters and a simpler one involving only two) are proposed that build on prior work. These frameworks are used to characterize the group's technology demonstrations. The insights gleaned from each analysis are discussed in the context of the needs and technology usage assumptions shared by the group, along with the costbenefit tradeoff of applying such frameworks.

Breaking the Human-Robot Deadlock: Surpassing Shared Control Performance Limits with Sparse Human-Robot Interaction

“…According to [32], LOA/FA is typically implemented in three ways 1 : adaptive autonomy, where the machine determines the level of assistance [39]; adjustable autonomy, where the human determines the level of assistance (sometimes called supervisory control [41]); or mixed-initiative control, where the human and machine deliberate to determine the level of assistance [44]. Sometimes, the term sliding autonomy is used to indicate one or a mixture of the three above [21].…”