Development of a low-cost IoT and machine learning-based air quality monitoring system for dairy barns

Advances in Internet of Things (IoT) and machine learning have enabled the development of cost-effective systems for real-time air quality monitoring in livestock environments. In this study, a low-cost Internet of IoT based device that can be used for air quality monitoring and evaluation in dairy barns was developed. Various machine learning algorithms were used to model the sensor data. The BME 680 air quality sensor data were collected in the ThingSpeak cloud database, and subsequent analysis was conducted. The performance of Random Forest (RF), Support Vector Regression (SVR), Gradient Boosting (GB), k-nearest neighbors (KNNs), Decision Trees (DTs), and Artificial Neural Networks (ANNs) algorithms was tested. The RF and GB algorithms yielded the best results in estimating air quality index values. However, it was stated that the performance of the ANN algorithm should be re-evaluated using large-volume datasets. In addition, it was emphasized that incorporating additional environmental parameters could improve the system’s performance. As a result, it was shown that low-cost digital agricultural applications can be developed using the capabilities of current technology.