Path planning for uncertainty reduced monitoring

For automating the monitoring works of a Remotely Operated Underwater Vehicle (ROV), we developed a path planning algorithm, which generates a efficient monitoring path for a semi-autonomous ROV. Firstly we categorized five typed sensor information in 2D Euclidean space, and defined the certainty for the sensor information and its space. Moreover, we defined the reliability function for comparing the superiority of various paths. Then, a modified genetic algorithm is used as a path planning algorithm. Two predefined paths and three random paths in a certainty space are compared with the paths generated by the planning algorithm. The experimental results showed that all generated paths by the path planning with genetic algorithm are superior than any other compared paths.