A comparative experimental study of two multi-robot olfaction methods: Towards locating time-varying indoor pollutant sources
Mingrui Jiang,Yu Liao,Xun Guo,Hao Cai,Wenqing Jiang,Zhou Yang,Fei Li,Fei Liu
DOI: https://doi.org/10.1016/j.buildenv.2021.108560
IF: 7.093
2022-01-01
Building and Environment
Abstract:Locating the source of indoor pollutants/hazardous substances is essential to ensure the health and safety of indoor personnel. The existing studies on source localization using robots rarely focus on time-varying sources whose release rate changes with time. However, in practical applications, time-varying sources are ubiquitous, and it is usually more difficult to locate a time-varying source than a constant source with a constant release rate. This study aims to locate time-varying sources in a real-world indoor environment by using our two newly proposed two multi-robot olfaction methods, namely, the improved particle swarm optimization (IPSO) method and the improved whale swarm algorithm (IWOA) method. To this end, we developed a source localization system composed of three robots and compiled the IPSO and IWOA methods into executable programs. This study conducted 160 experiments for four scenarios, which included two time-varying source types (periodic source and decaying source) and two source locations. In each scenario, 20 independent experiments were conducted for each method. For all four scenarios, the mean success rates of the IPSO and IWOA methods were 80% and 91%, respectively, indicating that both methods have the potential to locate indoor time-varying sources. In terms of the success rate, locating efficiency and locating accuracy, the overall performance of the IWOA method was better than that of the IPSO method. The advantages of the IWOA method were its success rate and locating accuracy, while the disadvantage was its locating efficiency.
engineering, environmental,construction & building technology, civil