Operator scheduling strategies in supervisory control of multiple UAVs

Abstract

The application of network centric operations to time-constrained command and control environments will mean that human operators will be increasingly responsible for multiple simultaneous supervisory control tasks. One such futuristic application will be the control of multiple unmanned aerial vehicles (UAVs) by a single operator. To achieve such performance in complex, time critical, and high risk settings, automated systems will be required both to guarantee rapid system response as well as manageable workload for operators. Through the development of a simulation test bed for human supervisory control of multiple independent UAVs by a single operator, this paper presents recent efforts to investigate workload mitigation strategies as a function of increasing automation. A humanin-the-loop experiment revealed that under low workload conditions, operators' cognitive strategies were relatively robust across increasing levels of automated decision support. However, when provided with explicit automated recommendations and with the ability to negotiate with external agencies for delays in arrival times for targets, operators inappropriately fixated on the need to globally optimize their schedules. In addition, without explicit visual representation of uncertainty, operators tended to treated all probabilities uniformly. This study also revealed that operators that reached cognitive saturation adapted two very distinct management strategies, which led to varying degrees of success. Lastly, operators with management-by-exception decision support exhibited evidence of automation bias.Keywords: multiple unmanned aerial vehicles, human supervisory control, levels of automation * Corresponding author, 77Massachusetts Ave, 33-305, Cambridge, MA 02139, missyc@mit.edu, 1-617-252-1512, 1-617-253-4196 (fax). IntroductionHuman supervisory control (HSC) occurs when a human operator monitors a complex system and intermittently executes some level of control on the system though some automated agent. Because HSC is not typically a continuous task, HSC operators can typically time-share cognitive resources with other tasks. This task time-sharing is critical for the success of network-centric operations (NCO) which require operators to process information from a variety of sources, often within a short period of time. One such domain where HSC and NCO will intersect in the future will be in unmanned vehicle operations. While many operators are presently needed to control a single unmanned aerial vehicle (UAV), as technology and autonomous control improve, automation will handle lower level tasks, thus enabling the supervisory controller more time to control a greater number of UAVs. The key to achieving this one-controlling-many goal will be the development of automated control architectures that compliment human decision-making processes in time-critical environments. A significant cognitive control issue for humans managing multiple vehicles is that they must synthesize voluminous data from a network of sensors and veh...

Levels of Automation in Human Factors Models for Automation Design: Why We Might Consider Throwing the Baby Out With the Bathwater

“…According to Onnasch et al (2014, p. 485), only four of the 18 studies included non-student participants; one of which employed military personnel not engaged in tasks pertaining to their expertise (i.e., Calhoun et al 2009). Similarly, Cummings and Mitchell (2007) recruited active duty military personnel with overlapping subject-matter expertise with respect to an anticipated future-world experimental task (Cummings, personal communication, March 31, 2017). The participants in Metzger and Parasuraman (2005) and Sarter and Schroeder (2001) were en route controllers and commercial aircraft pilots, respectively.…”

Identifying Predictive Metrics for Supervisory Control of Multiple Robots

“…Adherence to multiple principles are required to make this possible, including offloading low-level control of the robots to automation [3], [4], [5], [6], ensuring that the automation is reliable [7], and improving interface technologies (e.g. [8], [9]).…”

Intuitive Cognition and Models of Human–Automation Interaction

“…Onnasch et al (2014) mention six representative studies that show a greater negative impact of automation failure on performance during the decision, Stage 3, than during the analysis, Stage 2. However, only three of these studies actually lend themselves to the kind of analysis that I wish to do here: Sarter and Schroeder (2001), Cummings and Mitchell (2007), and Rovira, McGarry, and Parasuraman (2007). Sarter and Schroeder (2001).…”

“…Indeed, Cummings and Mitchell (2007) stated that the advantage of the passive display "is schedule integration through a representation aiding graphical format which allows users to perceptually assimilate multivariate pieces of information quickly instead of requiring them to use working memory." (p. 342).…”

“…The higher level of performance obtained with the status display (Stage 2) relative to the command display (Stage 3)-the so-called critical boundary-was likely due to the differential engagement of intuitive versus analytical cognition, respectively. Cummings and Mitchell (2007). Participants had to monitor four unmanned aerial vehicles (UAVs), replan aspects of the mission, and manually execute actions (e.g., firing payloads) under high versus low levels of replanning (i.e., workload).…”