IEEE logo

Kernel-Functions-Based Models for Acoustic Localization of Underwater Vehicles

Breno C. Pinheiro1, Ubirajara F. Moreno1, João T. B. de Sousa2 O.C. Rodríguez3,
1Department of Automation and System, Federal University of Santa Catarina, Florianópolis, Santa Catarina Brazil
2Underwater Systems and Technologies Laboratory (USTL), Faculty of Engineering, University of Porto, Porto, Portugal.
3LARSyS, University of Algarve, 8005-139 Faro, Portugal

Comments: download pdf
Ref.: IEEE Journal of Oceanic Engineering Vol.42(3), pp.603-618

This paper proposes a novel design for the localization system of autonomous underwater vehicles (AUV) using acoustic signals. The solution presented exploits models based on kernel functions with two main purposes: i) to reject outliers; ii) to correct or improve accuracy of measurements. The localization system discussed is based on well-established techniques such as Support Vector Data Description (SVDD) and AutoAssociative Kernel Regression (AAKR) derived from machine learning theory that utilizes heuristic models for classification and regression tasks respectively. By coupling the algorithm to the navigation system, we seek to reduce the sensitivity of the localization scheme to the reflected acoustic waves or fluctuations of underwater channel properties without modifying the solution used for data fusion or overloading the algorithm embedded in the vehicle. Data collected in the field with a light underwater vehicle (LAUV) was used to demonstrate the advantages of the proposed approach.

“©1992-2014 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for  advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.”