The initial monitoring issues apparent within the UV-Robotics domain can be divided into the following four categories:
Comparing these to our ontology in Section 4, the first two categories involve the general alert types plan constraint violated and constraint violation projected. However, they exist at different levels of abstraction and often have different temporal impact and associated monitoring requirements. The third category cleanly fits the adversarial activity detected alert type and triggers alerts for both autonomous control and user reporting. The fourth category is essential both to team-based automated operation and effective user interaction, and involves policy constraint violated alerts, reporting requirement alerts, and system problem detected alerts.