Localizing particulate matter sources in indoor environments with weak airflow: An experimental study using swarm intelligence methods

• Published in: Sustainable cities and society 2024-11, Vol.115, p.105820, Article 105820
• Main Authors: Li, Zhenfeng, Bao, Gaogui, Mao, Hongyi, Jiang, Mingrui, Shi, Yue, Peng, Fusheng, Cai, Hao
• Format: Article
• Language: English
• Subjects: Different types of pollutants; Indoor environment with weak airflow; Multi-robot source localization; Particulate matter (PM); Unknown-height sources
• ISSN: 2210-6707
• DOI: 10.1016/j.scs.2024.105820

•Shift from gaseous to PM source localization in weak airflow indoors.•WOA_3D outperformed PSO_3D in localizing unknown-height PM sources.•Higher success achieved in localizing elevated PM sources.•Identified PM settling as a major source localization challenge. In indoor environments with weak airflow, advective transport diminishes while turbulent dispersion prevails. The absence of usable airflow data and the unpredictability of turbulence-led pollutant dispersion pose significant challenges to mobile robots in localizing pollution sources. This study advances the field by shifting the focus from gases to particulate matter (PM) in such environments. An initial detailed examination of airflow dynamics and PM behavior under weak airflow conditions laid the foundation for extensive 3D source localization experiments using two methodologies: the Whale Optimization Algorithm (WOA_3D) and Particle Swarm Optimization (PSO_3D), executed through a specialized multi-robot system. Evaluation across 11 distinct scenarios and 165 trials showcased the methods’ adaptability to different source heights, PM sizes, and the transition from gaseous to particulate pollutants. WOA_3D notably excelled in localizing PM2.5 sources at 1.35 m heights, achieving a 73.3% success rate and proving consistently effective across various elevations. However, its efficacy waned with larger PM sizes, and it was less effective in transitioning from ethanol vapor to PM source localization. These results highlight the importance of addressing PM settling to improve PM localization strategies, effectively combining insights into PM behavior’s complexities with methodological progress.