Predictive Path Loss Modeling for Wireless Sensor Networks in Structured Mango Orchards: Using Machine Learning Models

Analysis of radio wave propagation for computing the path loss is vital in designing the wireless sensor networks for precision agriculture. The accurate prediction of the path loss helps us to optimize the wireless sensor networks even in dense vegetation. Therefore, by employing the suitable machine learning models the accuracy of the path loss predictions can be comprehended. The selection of relevant features is the key for the application of machine learning to predict the path loss. In this paper, these features are divided as system parameters and environmental parameters. By giving these features as input parameters to different machine learning models the path loss is predicted and compared with the actual measured values. In addition, the performance metric of the several machine learning models has been compared. Among all Machine learning models, it is found that the predictions of XGBoost is more accurate for structured vegetation orchards.