Evaluation of MLP and LSTM ANNs for Signal Prediction in a Tunnel at 5.8 GHz Using Measurement Data

This article evaluates the performance of two artificial neural network (ANN) techniques, multilayer perceptron (MLP) and long short-term memory (LSTM), in predicting the level of mobile radio signals inside a tunnel environment, with data collected at the 5.8 GHz frequency, mainly relevant for communication with and between vehicles. This study investigates the effectiveness of ANN-based models compared to a traditional reference model, log-distance (LD), for predicting received signal level. By analyzing the results, it was possible to identify the relevant factors of the MLP and LSTM networks that must be considered when building such models.