IoT-Enabled Air Purifiers with Real-Time Quality Monitoring

The aim of the research is to develop an Internet of Things (IoT)-based air purifier that uses Convolutional Neural Networks (CNN) to detect air quality in real-time. It will then compare the performance of the devices using Support Vector Machines (SVM) based on energy efficiency and air quality improvement. Materials and Methods: There are two groups utilized in this research work: Group 1 utilizes 26 instances of SVM to simulate IoT-based air cleaners, and Group 2 uses 26 examples of CNN for simulating IoT-based air cleaners. PM2.5, PM10, CO2, temperature, and relative humidity were read through IoT environmental sensors, whose readings were made available. The training and tests were conducted in Python TensorFlow, Scikit-Learn, and monitoring in real-time was conducted on Google Firebase and AWS IoT Core. 95% is the confidence level, 80% is the G Power value, and 0.05% is the threshold. Results: The CNN-based air purifier achieved 92.4% accuracy, greatly outperforming SVM at an accuracy of 85.7%. Despite the CNN-based model detecting pollutants 24% faster than SVM, energy consumption was decreased by 18%. With p=0.0056 statistical significance, the greatest reduction in air pollution was seen when PM2.5 was decreased by 65%. Conclusion: In contrast to SVM-based models, it has been shown that air purifiers backed by IoT using CNN tremendously enhance energy efficiency as well as the observation of air purity, thus being suitable for environmental surveillance and control against pollution.