Unit:
Κατεύθυνση Smart Telecom and Sensing Networks
(SMARTNET)Πληροφορική
Author:
Koc Muhammed Fatih
Supervisors info:
Dimitris Syvridis, Professor, NATIONAL AND KAPODISTRIAN UNIVERSITY OF ATHENS, SCHOOL OF SCIENCE
DEPARTMENT OF INFORMATICS AND TELECOMMUNICATION
Original Title:
MSc THESIS - Occupancy Detection in Indoor Environments Based on Wi-Fi Measurements and Machine Learning Methods
Translated title:
MSc THESIS - Occupancy Detection in Indoor Environments Based on Wi-Fi Measurements and Machine Learning Methods
Summary:
Mobile devices are currently significantly becoming part of our daily lives due to the wireless communication capabilities that have enabled a series of high-level services. These Wi-Fi equipment are continuously sending packets stated as probe requests that can be captured using wireless sniffers. In this thesis, we tried to solve the problem of exploiting such a methodology to complete occupancy estimation by considering how many people exist in a specific space. At first, we discussed collecting Wi-Fi probe request packets using the Raspberry Pi device and analysing them with packet analyzer tools. We operated data collection in different environments, ranges in different level densities and used the mobile camera as the ground truth value. Afterwards, we represented how we can use MAC addresses and power level information for indoor prediction in the proposed linear ridge regression model using different approaches. We introduced a cheap and precise occupancy estimation model based on the capture of Wi-Fi frames user’s devices. The model is applied on low-cost hardware and utilized a supervised learning model to fit different environments. The experiments of such indoor estimations have been implemented in different scenarios to demonstrate the validity of the proposed solution and evaluate its results of it. The outcomes specify that mobile devices have good potential for predicting of the number of people in the space.
Main subject category:
Technology - Computer science
Keywords:
IEEE 802.11, Probe Request, MAC Address, Crowd Counting
M_Fatih_KOC_MSC_Thesis.pdf (1 MB)
Open in new window
master_thesis-main.zip
2 MB
File access is restricted.
