Determine activity based on the classified identity of users by using Wi-Fi monitoring

Abstract

The Wi-Fi technologies are used in everyday life on numerous applications that detect the crowd information for commercial, security and other reasons. The Wi-Fi monitoring can be used for tracking people when they are moving along different access points. The results from the Wi-Fi monitoring can provide the location of the users in an area and therefore, useful information can be extracted. The goal of this project is to recognize the activity of different users for different sessions of a Wi-Fi network. The Wi- Fi dataset that is used, is acquired from the Wi-Fi network of the Delft University of Technology (TU Delft). Initially, the estimation of the users’ occupation is determined with the use of a Markov model with the information that is derived from the Wi-Fi dataset. Their possible identity is used, in order to estimate the activity that a user is probably doing at a specific location of the research area. The results on the use of the research area, are calculated and visualised in different spatial levels, campus, building and floor level. The use of the building complex of the TU Delft Campus, is examined during irregular hours, to allow efficient real estate management and provide security solutions.