Autonomous Navigation of Underwater Vehicle with Static and Dynamic Obstacles using PTEM

Abstract

Mapping for navigation in underwater environments can be a challenge due to water clarity, bottom types, sonar resolution, clutter, and a host of other environmental conditions. Combining these issues with the localization challenges of operating without the use of GPS and relying on dead reckoning makes navigation of autonomous underwater vehicles extremely complex. This paper looks at the environment and obstacle representation for an autonomous underwater vehicle (AUV) in two scenarios - one where only static obstacles are present, and another where dynamic ones are introduced. A coordinate-based traditional method of supplying the obstacle’s x and y location is used as a baseline and the Probabilistic Threat Exposure Map (PTEM) method is implemented to improve the navigation of the AUV through various environments. Using computer simulations, the AUV navigates through three different maps both with and without the presence of a moving obstacle using both the coordinate-based traditional method of obstacle representation and the PTEM method. Performance metrics of position error, path length, map generation time, and total simulation time are evaluated; these results showed that the PTEM method always resulted in a quicker simulation time and matched or shortened the AUV’s path length.