Human Performance in Cyber-Vulnerable Highly Automated Unmanned Aerial Systems
Type of DegreePhD Dissertation
DepartmentIndustrial and Systems Engineering
MetadataShow full item record
Automation is ever present in today’s complex systems of systems. Originally created to ease workload and burden on the human operator and to increase safety by automating dangerous tasks, automation has become practically ubiquitous in almost every human-machine system today. It is present in transportation systems, the industrial complex, and in military applications. While automation is capable of reducing human workload in the human-machine team, and, in some cases, of removing the human from the system all together, it has also been shown to degrade human performance in monitoring and supervising tasks. In some systems, the supervisory role of the human has shifted from multiple humans monitoring a single system to a single human monitoring multiple systems. Today’s increased computing power, optimized algorithms, and machine learning techniques have contributed to the ongoing trend of humans monitoring multiple systems. This may soon be prevalent in military unmanned aerial systems (UAS). These systems use both wired and wireless networks to communicate and to send and receive information to control the system. If not meticulously protected, these systems may be susceptible to cyber intrusions ranging from an adversary monitoring performance to degradation of the system to a point of failure. Humans rely on timely and appropriate indicators from cyber-physical systems to protect these systems. This dissertation examines human performance during the operation of a generalized UAS control station and user interface. This includes consideration of the impact of disparate operator backgrounds on detecting faults and performance measures when controlling multiple UAS using a common user interface as well as the impact of task loading, in the form of in-the-loop or out-of-the-loop participation. Additionally, the work compares two common UAS control station consoles from an ergonomics perspective, an important issue due to the prevalence of musculoskeletal disorders among pilots and military personnel and the resultant fatigue and discomfort that could further degrade operator performance in a cyber-compromised environment. The results of the first study found that the newer ground control station console, MD-4 Mission Control Station (MCS), reduced head-tilt extension angles in upward looking tasks and increased flexion in downward looking tasks resulting in improved lower gaze angles. However, there is still room to improve the postures and gaze angles that must be sustained. The data collection methods and analysis used to compare the two control station consoles could be used to evaluate other existing control stations and inform future ground control station designs for UAS currently being developed. The second study, which focused on cyber events and their impact on human performance, found that background was significant, with non-gamers worse than both aviators and gamers, while aviator and gamer performance metrics were generally comparable. Regardless of background, cyber events required significantly more time to respond. Task loading further impacted the accuracy of response to cyber events. These results could inform recruitment of future UAS operators focusing on certain experience backgrounds over others. They also highlight the need to further explore cueing of cyber events to operators and a better understanding of cyber events’ impact on mission accomplishment. Finally, the research could inform a conversation on the Navy’s current requirement for manned flight training for Group 4 and Group 5 UAS operators.
- Stutts Disseration.pdf