Point features extraction: towards slam for an autonomous underwater vehicle

Abstract

Simultaneous Localisation and Mapping (SLAM) is a process by which a mobile robot maps the environment and concurrently localises itself within the map. Feature extraction is a technique by which sensor data is processed to obtain well defined entities (features) which are recognisable and can be repeatedly detected. These features are then used to aid navigation. In this paper, Mechanically Scanned Imaging Sonar (MSIS) is used to perform scans of the environment. The information returned is then used to detect point features from data collected in a swimming pool. Artificial landmarks were introduced into the environment to obtain identifiable and stable features. This work is part of the authors efforts to develop a SLAM system to be utilised in an Autonomous Underwater Vehicle (AUV).