Intelligent vector-based path following guidance law for unmanned surface vehicles

Abstract

This study proposes a significant improvement on the classical vector field guidance law. Classical vector field method results in imaginary number some cases. In numerical calculation, imaginary part of the number ignores to obtain well-posed solution. However, negligence of imaginary part negatively affects the optimization workflow and it leads to larger cross tracking errors. In this study, classical vector field method was modified to obtain real numbers in all cases. Owing to the modified new method, cross tracking errors were diminished by 10% compared to the traditional vector field method. Second contribution of this study is to integrate the surrogate optimization technique to the developed robust vector field method in order to tune parameters of the guidance law. To the best of our knowledge, we are the first to combine surrogate optimization method with vector field path following algorithm. Vector field path following algorithm requires optimizing four important key parameters. While the unmanned surface vehicle performs the multi-command tasks, determining the optimum values of these four key parameters for each mission increases the computational costs significantly. Since the surrogate optimization method is more suitable for the time-consuming objective functions, it is preferred instead of the most popular optimization technique, genetic algorithm. Surrogate optimization method integrated robust vector field path following algorithm determines optimum parameters approximately 10 times faster than the genetic algorithm integrated vector field method. In addition to the time advantage of the developed model, the proposed method provides a navigation that causes less cross tracking errors compared to classical technique.