Using Neural Networks for RSSI Location Estimation in LoRa Networks

Localize objects in IoT applications is a very popular trend nowadays. In this work, we will present a neural network trained to estimate the location of an object in (Latitude and Longitude) using RSSI measurements through LoRa messages. The results of our work proving that neural networks can be used to to predict the location of objects in this kind of networks as our best model achieved average error on the test set 0.34 km, which is better from the average error without using neural networks.