Entropic skill assessment of unmanned aerial systems (Uas) operators

Abstract

Large-scale distributed training exercises involve many trainees at various stages of their training maturity and at various levels of skill. Problems arise in large-scale exercises when less mature or lower-skilled trainees are exposed to training scenarios that are too advanced or too complex for their level of training maturity. These trainees are more likely to fail the mission they are given in the training scenario, thus reducing the benefits of training, leading to frustration in the trainee or even disrupting the training exercise. We present a methodology for automated skill assessment using entropy measures that form the core of a battery of automated assessment algorithms. As illustrated in a case study, in which subjects performed a reconnaissance task in a simulated unmanned aerial system environment, this methodology achieves high accuracy levels of skill assessment and has the added benefit of computational simplicity, allowing for real-time skill assessment of trainees.