LoRaWAN-enabled Smart Campus: The Dataset and a People Counter Use Case

In brief, we present a novel framework that counts the number of people based on the readings of the sensors inside closed rooms. These sensors are connected using LoRa network. First, we identify the transmission failures that cause missing values in the dataset and handle such missing values. In addition, we validate the efficiency of handling the missing values using a long short-term memory (LSTM) architecture that forecast future values based on the handled missing values. Finally, we propose a neural network that formulate the number of people based on the readings. Simulation results show our model achieves an accuracy of 95% in predicting the number of people. We believe that our work would have a significant impact on the road of IoT and wide area network research in wireless communications. Our work could benefit lockdown restrictions during pandemic eras and smart air ventilation systems in smart residual building.