Indoor infection risk assessment through data evaluation of low-cost aerosol sensors using AI methods
Funding: Ministry of Science, Education and Culture, Rhineland-Palatinate
Regarding the SARS-CoV-2 pathogen, exhaled aerosol particles from infected individuals have been identified as the main transmission risk, especially indoors. In this regard, active ventilation in enclosed rooms such as classrooms or lecture halls can reduce spreading of potentially infectious particles. The resulting reduced concentration of pathogenic aerosols thereby minimizes the risk of infection. However, such novel ventilation concepts require a suitable control system to adjust the ventilation volume flow precisely and dynamically according to the current aerosol concentration. The control can be realized by online concentration measurement with suitable aerosol sensors. Here, the number and size distribution as well as the aggregate state of the aerosol particles determine the transmission and infection risk.
In this research project, the collected measurement data of aerosol sensors, which are developed for the detection of potentially infectious particles in another project at the institute, are evaluated with AI methods. For this, raw data from the aerosol sensors are combined with values from laboratory measurement devices to train a neural network to detect relevant particle sizes and characteristics. Thus, data on the actual concentration of potentially infectious particles can be collected and appropriate countermeasures can be initiated through adapted ventilation.