Probability-PSO Algorithm for Multi-robot Based Odor Source Localization in Ventilated Indoor Environments

Influenced by turbulence, the practical odor plume in ventilated indoor environments is fluctuant and intermittent. Moreover, the bigger eddies can easily lead to longtime local maxima. All these make the odor source localization quite complicated. A Probability Particle Swarm Optimization (P-PSO) algorithm is proposed for multi-robot based odor source localization in ventilated indoor environments. The P-PSO algorithm uses probability to express fitness function. In this paper, the odor source probability, which is estimated by Bayesian inference and variable-universe fuzzy inference, is taken as the fitness function. To validate the proposed search strategy, four different odor plumes corresponding to the real boundary conditions of an indoor environment are set up. Considering the response and recovery time of actual odor sensors, a second-order sensor model is built. Simulation results demonstrate the feasibility and advantage of the proposed P-PSO algorithm for multi-robot based odor source localization in ventilated indoor environments.