Finding Odours Across Large Search Spaces: A Particle Swarm-Based Approach

Abstract

This paper proposes an evolutionary-based search algorithm to find odour sources with robot communities across large search spaces. The characteristics of outdoor odour plumes and the main problems in detecting and finding them in real environments are described. An artificial olfaction system designed to carry out olfaction-based mobile robot experiments in realistic conditions is shown. This olfaction system is composed by intelligent gas sensing nostrils and a directional thermal anemometer. The searching algorithm proposed is inspired in the particle swarm optimization (PSO) method. This algorithm allows coordinating the movements of multiple robots searching for odour sources across large search spaces. The paper describes the algorithm and compares its performance against other searching strategies.