Persons counter through Wi-Fi’s passive sniffing for IoT

摘要

The objective of this work is to assess the use of MACs addresses as an effective counter of flows of people, with the final aim of generating time-series of users of public infrastructure in the context of the IoT. We present the results of a 35 days measuring campaign through the use passive sniffing of Wi-Fi devices. Essential data manipulation is performed on a Linux box through traditional commands. The seasonal component of the time series is analyzed. Thus, it can be modeled with an Auto-Regressive 1 model. Besides, we compare time series from different sensors, and results are similar and can complement each other to generate a robust sensor of flow of vehicles and persons. We have performed initial steps towards a pattern recognition, at a global and individual scale, of weekly behavior through a graphical representation and period detection.