Population Estimation Using Wi-Fi’s Received Signal Strength Indicator Based on Artificial Neural Network

Abstract

The development of population estimation using three (3) constructed received signal strength indicator (RSSI) acquisition devices with NodeMCU ESP8266 as the brain for data receiving and a Wi-Fi transmitter – all channeled into ThingSpeak for monitoring RSSI data and deployed into a designed graphical user interface (GUI) built and trained on MATLAB was demonstrated in this paper. The developed system considered a controlled indoor environment capable of predicting and estimating the number of people when moving and stationary. Based on the results of the training, validation, and testing for the two cases, an overall mean squared error of 1.36337 for moving with an overall response R-value of 0.87995 based on 125 hidden layers and 0.272564 for stationary with an overall response R-value of 0.98592 based on 95 hidden layers were obtained. The numerical results show that the model based on RSSI of Wi-Fi technology can classify the number of people inside the laboratory room from zero (vacant) up to 10 students.