Search

In term of motion planning for robots, several solutions are possible: grid-based search [1], interval-based search [2], geometric algorithms and potential fields [3]. However, potential fields offer a computational efficient way to generate a desired behavior for robots. However, a principal limit of potential fields is that they deal only with repulsion and attraction. To extend it to perpendicular, tangential and uniform fields, we base our paper on the works of S. Schmitt [4] and R. Arkin [3]. Our contribution consist in developing a method to construct complex vector fields, which are a linear combination of primitive fields. It also explains how to implement this method on a robot using the middleware ROS (Robot Operating System) with any controller.

A marine robot, especially a sailing boat robot, is exposed to a dynamic environment. This paper presents a simple and efficient obstacle avoidance control algorithm. The presented control method uses vector fields to regulate the marine robot.