Using genetic algorithms to fix a route for an Unmanned Surface Vehicle

Abstract

The paper reports the research whose the goal was to test whether genetic algorithms are able to effectively and quickly determine a route for an Unmanned Surface Vehicle (USV). In a normal working mode, the vehicle is manually controlled from Mobile Command Center (MCC) located on the seashore or on a ship. However, when a radio-connection between the vehicle and the MCC is disrupted, the USV has to independently reach a destination point, along a path fixed by a Route Fixing System (RFS). In the paper, a solution is presented in which the role of RFS is played by a Canonical Genetic Algorithm (CGA). Effectiveness of this algorithm was tested experimentally on artificial images imitating a sea chart. Results of the experiments are presented at the end of the paper.