Adaptive ventilation based on cough detection and localization for airborne infection control in vehicles.

Cough-generated pathogen-laden aerosols can lead to airborne transmission of infectious diseases inside vehicle cabins, where occupants share a confined space. Ventilation is the most effective approach for mitigating the risk of airborne infections in vehicles. However, the effectiveness of ventilation strongly depends on the location of the index occupant. To enhance airborne infection control, this study developed an adaptive ventilation method that detects and localizes the coughing occupant and adjusts the airflow distribution of the cabin ventilation system accordingly. A system integrating a three-microphone array, seat pressure sensors, and machine learning was built to detect and localize coughs in real time. Upon cough localization, ventilation airflow was adjusted through the car's vents to achieve adaptive ventilation. Effective ventilation settings for each index occupant location were identified via tracer aerosol experiments. The proposed method was then demonstrated using airborne bacteria experiments with a human participant coughing and a heated manikin receptor in the car. The results show that the cough detection and localization system achieved 98% accuracy and 95% accuracy, respectively, in the experimental vehicle. The demonstration experiments show that when the human participant in the driver seat coughed while releasing Lactobacillus bulgaricus, the adaptive ventilation reduced the average airborne Lactobacillus bulgaricus concentrations in the breathing zone of the receptor manikin by 84%, 71%, and 79% for the passenger, rear-left, and rear-right seats, respectively, compared with the common ventilation typically used in the car.