Concurrent mapping and localization with FLS

Abstract

The topic of concurrent mapping and localization (CML) using a forward look sonar (FLS) is addressed. In many AUV applications, artificial beacons are employed for long duration positioning, or Global Positioning System (GPS) data are used to provide navigation resets. There are numerous missions where these constraints are undesirable and it is the aim of CML to provide relative position by extracting information from the environment through which it is moving. A technique for CML using a multibeam FLS is presented. The method is based on forming objects based on compact regions of high-level bottom backscatter, extracting features from the objects to identify/label them to aid in the data association process, and using the objects in an extended Kalman filter (EKF). A simplified approach to the EKF is used, whereby a new and separate filter is initiated for each new object encountered. Vehicle position is then derived as a weighted sum of the outputs from the individual filters. This approach has been developed and tested using data collected with an advanced FLS referred to as the high resolution array (HRA).