An Experimental and Numerical Study on a Multi-Robot Source Localization Method Independent of Airflow Information in Dynamic Indoor Environments

Qilin Feng,Hao Cai,Yibin Yang,Jiheng Xu,Mingrui Jiang,Fei Li,Xianting Li,Chenchu Yan

DOI: https://doi.org/10.1016/j.scs.2019.101897

IF: 11.7

2019-01-01

Sustainable Cities and Society

Abstract:To locate contaminant sources in dynamic indoor environments, this study presents an improved particle swarm optimization (IPSO) method independent of airflow information and validates the method by combining robot experiments with numerical simulations. The experimental study was first conducted by using three mobile robots to locate an ethanol source in a typical dynamic indoor environment with a fan swinging periodically from left to right. A total of 12 out of 15 experiments were successful, with a success rate of 80%, indicating that the method has a high success rate and strong robustness. Next, the experimental environment was further simulated by CFD, and numerical experiments were conducted. The results show that the success rate and the average number of steps from numerical experiments were consistent with those from robot experiments, indicating the feasibility of using numerical simulations. Finally, the IPSO method was numerically validated and compared with a standard PSO (SPSO) method and a modified PSO method with wind utilization II (WUII) in mixing ventilation (MV) and natural ventilation (NV) cases. With a similar average number of steps, the IPSO method achieved a higher success rate than the comparison methods, indicating the superiorly of the IPSO method.