Evolutionary path planning and navigation of autonomous underwater vehicles

Abstract

This paper presents a complete methodology for mission planning and navigation of autonomous underwater vehicles (AUVs) in ocean environment. Path planning near the ocean floor is accomplished via genetic algorithms and B-splines based on known data of the ocean floor. In addition, collision free navigation is achieved in unknown environments. Prior to vehicle's launch, a genetic algorithm based on ocean floor data and on mission restrictions calculates the optimal path. Once the path is calculated, the vehicle is navigated through the predefined path by a set of fuzzy controllers. A second evolutionary algorithm optimizes the membership functions of these controllers so as the vehicle has the minimum error through its course. Extensive simulations were performed in order to evaluate the methodology and the derived optimized controller.