Coordination of a school of robotic fish using nearest neighbour principles

Abstract

Autonomous Underwater Vehicles (AUVs) are Unmanned Underwater Vehicles (UUVs) that are able to function without direct control from a human operator. Consequently, they have a wide range of applications from scientific research of the oceans to military applications such as maritime surveillance. However, there is now the demand for AUVs to be operated within a multi-vehicle scenario to allow large areas of the ocean to be monitored simultaneously. However, in order for this to become a reality algorithms have to be created that ensure that a group of AUVs could be self-organising. Therefore, using a validated mathematical model of a biomimetic robotic fish (called RoboSalmon) and taking inspiration from nature, this paper outlines the implementation of co-ordination algorithms based upon the behavioural mechanisms exhibited by schools of fish to allow a group of AUVs to become self-organising. The algorithms implemented are based on two different methodologies known as the Discrete and Continuous Behavioral Zone methodologies. The results obtained demonstrated that although both methodologies result in the formation of a school structure, the results obtained from the Continuous Behavioral Zone (CBZ) methodology were more resilient to changes in parameters associated with school structures and therefore these algorithms provided the most effective way to allow a group of AUVs to be considered as self-organising.