An improved particle swarm optimization method for locating time-varying indoor particle sources

Area of Science:

• Environmental Science
• Robotics
• Chemical Engineering

Background:

• Indoor airborne particles pose health risks, necessitating effective source localization.
• Identifying time-varying particle sources is challenging due to complex transport and release dynamics.

Purpose of the Study:

• To develop an improved multi-robot olfactory search method for locating time-varying indoor particle sources.
• To address challenges in particle source identification caused by variable release rates and complex airflows.

Main Methods:

• Proposed an enhanced particle swarm optimization (PSO) algorithm incorporating an upwind term for multi-robot olfactory search.
• Simulated particle concentrations and air velocities using computational fluid dynamics (CFD) across four scenarios with different ventilation types and source characteristics.
• Validated CFD simulations with experimental data for accuracy.

Main Results:

• The improved PSO method successfully located periodic and decaying particle sources.
• Achieved source localization success rates exceeding 96%, significantly outperforming standard PSO and Wind Utilization II algorithms.
• Demonstrated effective source identification within approximately 55 seconds across various ventilation conditions.

Conclusions:

• The proposed multi-robot olfactory search method offers a highly effective solution for indoor particle source localization.
• The integration of air velocity data and an enhanced PSO algorithm improves accuracy and efficiency in complex indoor environments.
• This approach provides a robust tool for mitigating health risks associated with indoor airborne contaminants.