A Neural Network-based Path Loss Model for Bluetooth Transceivers

With the development of Bluetooth communication, wireless communication technology is being applied in various fields such as tracking the location of smart devices and building an Internet of Things environment. The received signal is affected by various factors such as weather or terrain located on the transmission/reception path. Therefore, estimating the degree of path loss using a mathematical propagation model involves considerable errors. In this paper, we propose the artificial neural network-based path loss models to accurately estimate path loss. As the model's performance evaluation metric, the mean absolute error (MAE), mean square error (MSE), root mean square error (RMSE), mean absolute perception error (MAPE), and R2 score are used. Performance comparisons with several neural network-based models were performed, and actual data were collected using Bluetooth devices for model training and performance evaluation. Among the various models, the proposed path loss model was the best because it achieved a mean absolute error of 3.71m.