CNN-Based Indoor Path Loss Modeling with Reconstruction of Input Images

Convolutional Neural Networks (CNNs) have shown surprisingly good performance in both classification and regression problems. Given a floor plan of a building and indoor measurement data of Wi-Fi received signal strength (RSS) at discrete locations, a CNN can be trained to approximate the underlying functions of path loss. In this article, we propose a novel CNN-based indoor path loss modeling approach. Based on a floor plan and measurement data, input images are generated for training a CNN, which can make predictions of the RSS of 5 GHz Wi-Fi in an indoor usage scenario. Experiment results show that CNNs can be used for indoor path loss modeling with the encouraging performance with the Root Mean Square Error (RMSE) of 3.9404 dBm and good generalization property.