Supervising UAVs : improving operator performance by optimizing the human factor

Tele-operated unmanned aerial vehicles (UAVs) have no operators on board and therefore enable extension of the present sensing and communication capabilities in civil and military missions, without unnecessarily endangering personnel or deploying expensive material. One should also realize that tele-operation from a control centre may result in new and formerly unknown human factors problems. For this reason, TNO Human Factors has been performing UAV related research for more than ten years, focusing on the identification of possible human factors problems in UAV supervisory control. The present paper gives an overview of exploratory human-machine interface optimization studies, and reports the findings to overcome specific problems that are inherent in remote camera and platform operation. Important points of departure for the studies are a UAV mission analysis and an assessed order of possible missions. Since important mission elements involve target location, classification (target acquisition), and damage assessment, operator-in- he-loop performance remains an important issue, depending on the level of automation. We acquired more knowledge on the functioning of the human operator dealing with remote sensing and control, with the emphasis on aspects related to automation and interface aspects. For automation this involved, among others, the distribution of tasks between the human operator and the unmanned system, automation and operator situation awareness, the effects of automation breakdowns and adaptive automation. The interface aspects mainly involved the viewing systems and control devices that form the interface between the remote operator and the UAV. This concerned, among others: image characteristics of the remote sensor, application and consequences of head-slaved viewing systems, effects of time delays, control device configurations and the application of force feedback. The paper discusses a series of studies that focuses on the negative effects of degraded information, and the possibilities to compensate these effects by innovative human- machine interface designs supporting the operator in charge. Important point of departure was that the improvements did not result in additional claims on the capacity of the up/down data-link. The applied techniques included the use of graphic overlays, ecological interface design, head-coupled control, and the use of prediction techniques. The results show that carefully designed human-machine interfaces are able to partially compensate specific image degradations while operator performance may significantly increase. Further experimental research focused on sub-optimal operator performance when UAV remote information is handled on board a moving platform, e.g., a helicopter. We tested methods to increase spatial awareness. The paper gives an overview of the studies and presents the main results.