A reactive and knowledge based autonomous underwater vehicle

An architecture of the high-level controller and a world model for a single-mission configurable autonomous underwater vehicle are presented. The model depicted is a reactive behavior layered for interaction with the user. A dedicated watch-dog layer monitors the whole structure to detect and solve deadlock problems. The internal world model (WM) is analyzed, and a classification of WM knowledge is introduced. Its integration in the architecture is studied, focusing on the knowledge based part of the WM and its role in the user interface. A case study using the concepts introduced is also presented