Evaluation and Calibration of a Low-cost Particle Sensor in Ambient Conditions Using Machine Learning Methods

Particle sensing technology has shown great potential for monitoring particulate matter (PM) with very few temporal and spatial restrictions because of its low-cost, compact size, and easy operation. However, the performance of low-cost sensors for PM monitoring in ambient conditions has not been thoroughly evaluated. Monitoring results by low-cost sensors are often questionable. In this study, a low-cost fine particle monitor (Plantower PMS 5003) was co-located with a reference instrument, named Synchronized Hybrid Ambient Real-time Particulate (SHARP) monitor, in Calgary Varsity air monitoring station from December 2018 to April 2019. The study evaluated the performance of this low-cost PM sensor in ambient conditions and calibrated its readings using simple linear regression (SLR), multiple linear regression (MLR), and two more powerful machine learning algorithms using random search techniques for the best model architectures. The two machine learning algorithms are XGBoost and feedforward neural network (NN). Field evaluation showed that the Pearson r between the low-cost sensor and the SHARP instrument was 0.78. Fligner and Killeen (F-K) test indicated a statistically significant difference between the variances of the PM_2.5values by the low-cost sensor and by the SHARP instrument. Large overestimations by the low-cost sensor before calibration were observed in the field and were believed to be caused by the variation of ambient relative humidity. The root mean square error (RMSE) was 9.93 when comparing the low-cost sensor with the SHARP instrument. The calibration by the feedforward NN had the smallest RMSE of 3.91 in the test dataset, compared to the calibrations by SLR (4.91), MLR (4.65), and XGBoost (4.19). After calibrations, the F-K test using the test dataset showed that the variances of the PM_2.5 values by the NN and the XGBoost and by the reference method were not statistically significantly different. From this study, we conclude that feedforward NN is a promising method to address the poor performance of the low-cost sensors for PM_2.5 monitoring. In addition, the random search method for hyperparameters was demonstrated to be an efficient approach for selecting the best model structure.颗粒物传感技术凭借低成本、体积小巧且操作简便的优势，在几乎无时空限制的颗粒物（PM）监测领域展现出巨大应用潜力。然而，当前用于环境颗粒物监测的低成本传感器的实际性能尚未得到全面评估，其监测结果往往可信度存疑。本研究于2018年12月至2019年4月期间，将一款低成本细颗粒物监测仪（Plantower PMS 5003）与参考设备——同步混合环境实时颗粒物（SHARP）监测仪，同步布设至卡尔加里Varsity空气监测站。研究评估了该低成本PM传感器在环境条件下的性能，并采用简单线性回归（SLR）、多元线性回归（MLR）以及两种借助随机搜索技术优化模型架构的高性能机器学习算法对其读数进行校准，这两种算法分别为极限梯度提升（XGBoost）与前馈神经网络（NN）。实地评估结果显示，未校准的低成本传感器与SHARP监测仪的皮尔逊相关系数（Pearson r）为0.78。弗利格纳尔-基林（F-K）检验表明，低成本传感器与SHARP监测仪测得的PM₂.₅数值方差存在统计学意义上的显著差异。实地观测发现，校准前低成本传感器存在显著的高估现象，该现象被认为由环境相对湿度的变化所致。未校准的低成本传感器与SHARP监测仪的均方根误差（RMSE）为9.93。在测试数据集上，前馈神经网络校准后的RMSE最低，为3.91，而简单线性回归、多元线性回归与XGBoost校准后的RMSE分别为4.91、4.65与4.19。校准后，基于测试数据集的F-K检验结果显示，前馈神经网络、XGBoost校准后的PM₂.₅数值方差与参考方法的方差无统计学意义上的显著差异。本研究结论表明，前馈神经网络是改善PM₂.₅监测用低成本传感器性能的有效方法；此外，基于随机搜索的超参数优化方法被证实为筛选最优模型架构的高效手段。