Effect of Stages and Levels of Automation and Reliability on Workload and Performance for Remotely Piloted Aircraft Operations.

摘要

This thesis investigates incorporating different stages and levels of automation with varying degrees of reliability into a remotely piloted aircraft (RPA) surveillance task in order to determine how automation implementation and reliability affect operator workload and system performance. The study uses IMPRINT discrete event simulation to evaluate three levels of reliability in twelve different baseline automation implementations within a remotely piloted vehicle task. Three stages and four levels are modeled, for a total of twelve combinations, along with a baseline task with no automation. The stages modeled are the information acquisition stage, the decision and action selection stage, and the action implementation stage, coupled with the automation recommendation level, the operator consent level, the operator veto level, and the fully automatic level. The reliability is assessed at 100%, with reduced reliabilities of 80%, 70%, and 60%. This study finds that stages of automation have greater impact on performance and the workload values than levels of automation. Automation with reduced reliability is found to have significantly reduced performance for all stages except the response stage models. However, reductions in reliability are found to have little impact on operator workload.