Source Localization in Dynamic Indoor Environments with Natural Ventilation: an Experimental Study of a Particle Swarm Optimization-Based Multi-Robot Olfaction Method

Qilin Feng,Canxin Zhang,Jingyu Lu,Hao Cai,Zhilong Chen,Yibin Yang,Fei Li,Xianting Li

DOI: https://doi.org/10.1016/j.buildenv.2019.106228

IF: 7.093

2019-01-01

Building and Environment

Abstract:Source localization is important for ensuring indoor air quality and indoor environmental safety. Most available studies on source localization have been conducted in steady indoor environments with mechanical ventilation, and only a very few studies have addressed the more challenging source localization problem in indoor environments with natural ventilation. To locate contaminant sources in indoor environments with natural ventilation, this study presents a multi-robot olfaction method (URPSO) based on an adapted particle swarm optimization (PSO) algorithm by adding an upwind term and a random disturbance term into the standard PSO algorithm. The effectiveness of the presented method was first validated by three robots in a natural ventilation environment created by opening a window. For two typical source locations in the downwind zone and the recirculation zone, 13 and 15 experiments out of 15 experiments were successful, with success rates of 86.7% and 100% and average steps for source localization of 28.7 and 18.4 steps, respectively. The URPSO method was further compared with the PSO and wind utilization II (WUII) methods for locating the source in an imitated natural wind environment produced by using a fan. For the URPSO, PSO and WUII methods, 14, 3 and 5 experiments out of 15 experiments were successful, with success rates of 93.3%, 20% and 33.3% and average steps of 34.1, 34.0 and 35.8 steps, respectively. The experimental results show that the presented method has a similar source localization efficiency but a much higher success rate than the compared methods.