Comparing Quantitative Model Predictions to Experimental Data in Multiple-UAV Flight Control

Abstract

Thirty-six licensed pilots from the University of Illinois Aviation Institute performed simulated military surveillance missions with one and/or two unmanned aerial vehicles (UAV). Pilots were responsible for navigating each UAV through a series of mission legs in one of the following conditions: 1) a baseline condition with all manual flight controls and visual displays; 2) an auditory offload condition that provided auto-alerts and other relevant information to the auditory channel; and 3) an automation condition that provided auto-pilot control of the UAV. Pilots were responsible for mission completion, target search, and systems monitoring. Results indicate that the two offloads are beneficial in reducing task interference and overall workload. Three theories, with corresponding workload models, were discussed in order to evaluate predicted pilot performance. Single channel theory was able to explain some of results in the baseline condition, while single resource theory and multiple resource theory were better able to explain reduced task interference in the automation and auditory conditions.