Cognitive knowledge representation under uncertainty for autonomous underwater vehicles

Abstract

This paper presents an early approach for marine cognitive robots, in order to incorporate uncertainty into an ontological representation of the world. The proposed system is based on a signal processing module and an ontology-based knowledge framework, which is queried and updated according to the processed sensor data. It has been successfully demonstrated post-processing data from a mission of NessieAUV at The Underwater Centre, in Fort William, west of Scotland. The system shows its ability to process sensor data, identify basic features (lines and circles) and populate the ontology model. Additionally, from the ontology side, the basic information are elaborated in order to arrive to more complex concepts, like pillars and crossbeams.