Localising odour source using multi-robot and anemotaxis-based particle swarm optimisation

Localising an odour source is of considerable importance in real-world applications. A modified particle swarm optimisation (PSO) algorithm was presented in this study as an effective cooperative strategy among robots to guide them to search for an odour source. In this algorithm, the whole search space is divided into several grid cells with each cell saving positions that have been detected by robots. In the phase of searching for an odour plume, each robot locating at a cell obtains a repulsive force proportional to the number of positions saved by this cell. To improve the efficiency of particles in traversing the plume, two cognitive factors of PSO are dynamically adjusted according to the effect of wind on self-cognition and social cognition of a particle. In addition, simulated annealing is incorporated into the update of a particle´s local leader to prevent premature convergence of the swarm. The proposed algorithm, implemented using MATLAB, was applied to localise the odour source in four typical simulation scenes formed by Fluent, and compared with previous methods. The experimental results show that the proposed algorithm can guide a multi-robot system to localise the odour source rapidly and accurately.