On the use of particle filters for terrain based navigation of sensor-limited AUVs

Different Terrain Based Navigation systems for underwater vehicles have already been presented, with experimentally validated results and consistent performance. However, these results are mostly based on the use of both high accuracy inertial navigation systems and high quality sonars. This article presents a study on Particle Filter algorithms that cope with peculiarities of Terrain Based Navigation for sensor limited systems. The focus is on the influence on several parameters, namely the process noise, the measurement noise and the number of the particles, and how these can improve the obtained results. Based on the results obtained by simulation, we present some conclusions relevant for the design of future implementation of the algorithms.