Swarm potential fields with internal agent states and collective behaviour

Abstract

Swarm robotics is a new and promising approach to the design and control of multi-agent robotic systems. In this paper we use a model for a system of self-propelled agents interacting via pairwise attractive and repulsive potentials. We develop a new potential field method using dynamic agent internal states, allowing the swarm agents' internal states to manipulate the potential field. This new method successfully solves a reactive path planning problem that cannot be solved using static potential fields due to local minima formation. Simulation results demonstrate the ability of a swarm of agents that use the model to perform reactive problem solving effectively using the collective behaviour of the entire swarm in a way that matches studies based on real animal group behaviour.