Accuracy assessment of air quality (PM2.5 and PM10) sensor on a drone affected by airflow

Abstract

With the recent case of forest fires from the region of Java, Indonesia, affecting Singapore and her closest neighbors, the attention on air quality has increased in the public domain. There were more reports in local newspapers in Singapore about the forest fires and how it could negatively affect our health if proper measures individually were not taken and ways to curb this problem. And with the rise of smartphone applications, the information of the Pollution Standards Index (PSI) can be easily made known to the public wherever they are. However, these PSI values were taken from multiple reporting stations in Singapore, which are averaged out. Hence, the accuracy of the values might be affected if a person is located further away from these stations or the air quality information is needed in a localized area. Henceforth, there were many suggestions on getting greater accuracy on air quality information, which can lead to better-informed decisions and a suggestion was to use a drone. The functions and availability of drones are varied and its potentiality in collecting air quality information will be explored here; hence, assessing the accuracy of putting sensors on the drone. And it is found that putting sensors in between adjacent rotors out of other variations give the most accurate readings. The accuracies are analyzed by the data given together with the analysis of variance method (ANOVA) and three factors were studied to test its impact on the particle count. Vertical variation was the only factor that had significant impact on the readings. Hence, it was concluded that in addition in placing the sensor at the best position, engineering a case that can deflect away the downstream airflow from the rotors would increase the accuracy of the readings.