Democratic Inspired Particle Swarm Optimization for Multi-Robot Exploration Task

Oussama Moslah (1) , Yassine Hachaïchi (2) , Younes Lahbib (1)

1 ENIT - Ecole Nationale d'Ingénieurs de Tunis
2 LR-LAMSIN-ENIT - Laboratoire de Modélisation Mathématique et Numérique dans les Sciences de l'Ingénieur [Tunis]

In this paper, we propose a new method for exploring an unknown environment with a team of homogeneous mobile robots. The goal of our approach is to minimize the exploration time. The challenge in multi-robot exploration is how to develop distributed algorithm to govern the colony of robots while choosing its new direction so that they simultaneously explore different regions. In this paper we use the extended version of Particle Swarm Optimization (PSO) to robotic applications, which is referred to as Robotic Particle Swarm Optimization (RPSO), a technique to compute robots new location. To better adapt this technique to the collective exploration problem, and maximize the exploring area, we propose a new method for computing PSO's global best parameter. Experiment results obtained in a simulated environment show that our new method of computing PSO's global best parameter increases the explored area with a shorter time convergence.

Mots clés: collective exploration; Multi-robot system; Swarm robotic; mobile robotics; SLAM; Particle Swarm Optimisation

Domaines: Mathématique discrète [cs.DM]; Complexité [cs.CC]

Oussama Moslah, Yassine Hachaïchi, Younes Lahbib. Democratic Inspired Particle Swarm Optimization for Multi-Robot Exploration Task. 2016. ⟨hal-01327367⟩