Flight testing of agent supervisory control on heterogeneous unmanned aerial system platforms

This article describes the execution of flight experiments with a cognitive agent architecture on two different airborne platforms; a fixed-wing unmanned aerial vehicle (UAV) and a vertical takeoff and landing (VTOL) multirotor aircraft. Two separate experimental reconnaissance missions were conducted in a single operator single UAV setup. The missions were designed to leverage the strengths of each platform. Both aircrafts were operated using the same human machine interface (HMI) and the same task-based guidance command style. A cognitive agent running onboard the systems employed different capabilities to fulfill the given tasks and to complete the mission in a highly automated fashion. It decided about the employment depending on the situation, either with or without interaction with the human operator. The system architecture proved to be functional under real flight conditions on heterogeneous platforms and allowed the operator to focus on the mission instead of low-level platform details. Yet it allowed access to those functionalities if necessary.