Acoustic based navigation of cooperative marine robots with advanced filter techniques

Abstract

In this paper, we discuss a possible navigation solution for two autonomous marine robots, of which one is submerged. The situation is a sub task of the overall mission scenario of a recent research project. The goal is to provide the submerged robot with an estimation of the positon of the surface unit relative to its own (unknown) position. On that base, a control algorithm can be developed, what we cite from literature. The navigation solution that we propose here is an enhancement of an older version we published before where we used a linear Kalman filter and ‘pseudo-linear’ measurements. Our new solution is based on the employment of an Unscented Kalman Filter, therefore the non-linear structure of system model and measurement model can be used. We will compare the new solution with the older one in MATLAB simulations and demonstrate that it leads to even better results.