Path planning for multiple marine vehicles

Motivated by increasingly complex and challenging missions at sea, there is widespread interest in the development of advanced systems for cooperative control of multiple autonomous marine vehicles. Central to the implementation of these systems is the availability of efficient algorithms for multiple vehicle path planning that can take explicitly into account the capabilities of each vehicle and existing environmental conditions. Examples include path planning to steer a group of marine vehicles and enable them to reach a specific target site simultaneously with a desired formation pattern, while avoiding inter-vehicle collisions, and online path replanning for a vehicle fleet upon detection of episodic events or obstacles. Multiple vehicle path planning methods build necessarily on key concepts and algorithms for single vehicle path following. However, they go one step further in that they must explicitly address such issues as inter-vehicle collision avoidance and simultaneous times of arrival. As such, they pose considerable challenges both from a theoretical and practical implementation standpoint. This paper is a brief survey of multiple vehicle path planning techniques. The exposition is focused on specific algorithms for path planning developed in the scope of research work in which the authors have participated. The algorithms make ample use of direct optimization methods that lead to efficient and fast techniques for path generation. The paper affords the reader a fast paced presentation of key algorithms that had their genesis in the aircraft field, discusses them critically, and suggests problems that warrant further consideration.

I. Introduction

There is increasing demand for the use of autonomous mobile robots, which are steadily becoming the tools par excellence for the execution of challenging missions in areas that are hard to access or place human lives at risk. Space, land, and marine robots are by now ubiquitous and hold promise to the development of networked systems to sample the environment at an unprecedented scale.