Air quality estimation using low-cost sensors is a pressing issue, with meteorological factors often causing measurement discrepancies. Hygroscopicity, arising from humidity’s interaction with particulates, leads to inaccurate PM concentration readings in laser-scattering low-cost PM sensors. Common remedies involve data removal during high relative humidity or reference station calibration, but these solutions are not always practical or accurate due to the localized nature of hygroscopicity. In this paper, the authors present an adaptive correction framework that dynamically models hygroscopicity effectively mitigating humidity’s impact on particle measurements. The framework exploits historical sensors’ data, providing real-time adaptability in any context without relying on reference data, thus improving air quality estimations from low-cost sensors.

使用低成本传感器估算空气质量是一个紧迫的问题，气象因素常常导致测量差异。湿度与颗粒物相互作用产生的吸湿性会导致激光散射低成本 PM 传感器的 PM 浓度读数不准确。常见的补救措施包括在高相对湿度或参考站校准期间删除数据，但由于吸湿性的局部性质，这些解决方案并不总是实用或准确。在本文中，作者提出了一种自适应校正框架，可以动态模拟吸湿性，有效减轻湿度对颗粒测量的影响。该框架利用历史传感器数据，在任何情况下提供实时适应性，而无需依赖参考数据，从而改进低成本传感器的空气质量估计。