Moursi Ahmed Samy, El-Fishawy Nawal, Djahel Soufiene, Shouman Marwa Ahmed
Computer Science and Engineering Department, Faculty of Electronic Engineering, Menoufia University, Menouf, 32952 Menoufia Governorate Egypt.
Department of Computing and Mathematics, Manchester Metropolitan University, Manchester, M15 6BH UK.
Complex Intell Systems. 2021;7(6):2923-2947. doi: 10.1007/s40747-021-00476-w. Epub 2021 Jul 29.
Air pollution is a major issue resulting from the excessive use of conventional energy sources in developing countries and worldwide. Particulate Matter less than 2.5 µm in diameter (PM) is the most dangerous air pollutant invading the human respiratory system and causing lung and heart diseases. Therefore, innovative air pollution forecasting methods and systems are required to reduce such risk. To that end, this paper proposes an Internet of Things (IoT) enabled system for monitoring and predicting PM concentration on both edge devices and the cloud. This system employs a hybrid prediction architecture using several Machine Learning (ML) algorithms hosted by Nonlinear AutoRegression with eXogenous input (NARX). It uses the past 24 h of PM, cumulated wind speed and cumulated rain hours to predict the next hour of PM. This system was tested on a PC to evaluate cloud prediction and a Raspberry P to evaluate edge devices' prediction. Such a system is essential, responding quickly to air pollution in remote areas with low bandwidth or no internet connection. The performance of our system was assessed using Root Mean Square Error (RMSE), Normalized Root Mean Square Error (NRMSE), coefficient of determination ( ), Index of Agreement (IA), and duration in seconds. The obtained results highlighted that NARX/LSTM achieved the highest and IA and the least RMSE and NRMSE, outperforming other previously proposed deep learning hybrid algorithms. In contrast, NARX/XGBRF achieved the best balance between accuracy and speed on the Raspberry P .
空气污染是发展中国家和全球范围内过度使用传统能源所导致的一个主要问题。直径小于2.5微米的颗粒物（PM）是侵入人类呼吸系统并导致肺部和心脏疾病的最危险空气污染物。因此，需要创新的空气污染预测方法和系统来降低此类风险。为此，本文提出了一种基于物联网（IoT）的系统，用于在边缘设备和云端监测和预测PM浓度。该系统采用混合预测架构，使用由带外部输入的非线性自回归（NARX）托管的几种机器学习（ML）算法。它利用过去24小时的PM、累积风速和累积降雨小时数来预测下一小时的PM。该系统在个人电脑上进行了测试以评估云端预测，并在树莓派上进行了测试以评估边缘设备的预测。这样的系统至关重要，能够对低带宽或无互联网连接的偏远地区的空气污染做出快速响应。我们的系统性能使用均方根误差（RMSE）、归一化均方根误差（NRMSE）、决定系数（ ）、一致性指数（IA）和以秒为单位的持续时间进行评估。所得结果表明，NARX/LSTM实现了最高的 和IA以及最小的RMSE和NRMSE，优于其他先前提出的深度学习混合算法。相比之下，NARX/XGBRF在树莓派上实现了准确性和速度之间的最佳平衡。
