Wi-Fi based people tracking in challenging environments

Khalili, Abdullah (2018) Wi-Fi based people tracking in challenging environments. Doctoral thesis, Staffordshire University.

Preview	Text (DPhil Thesis) Thesis.pdf - Submitted Version Available under License All Rights Reserved. Download (2MB) | Preview
	Text (Ethos agreement) EThOS-Deposit-Agreement.pdf - Other Restricted to Repository staff only Available under License All Rights Reserved. Download (233kB) | Request a copy

Abstract or description

People tracking is a key building block in many applications such as abnormal activity detection, gesture recognition, and elderly persons monitoring. Video-based systems have many limitations making them ineffective in many situations. Wi-Fi provides an easily accessible source of opportunity for people tracking that does not have the limitations of video-based systems. The system will detect, localise, and track people, based on the available Wi-Fi signals that are reflected from their bodies. Wi-Fi based systems still need to address some challenges in order to be able to operate in challenging environments. Some of these challenges include the detection of the weak signal, the detection of abrupt people motion, and the presence of multipath propagation. In this thesis, these three main challenges will be addressed.
Firstly, a weak signal detection method that uses the changes in the signals that are reflected from static objects, to improve the detection probability of weak signals that are reflected from the person’s body. Then, a deep learning based Wi-Fi localisation technique is proposed that significantly improves the runtime and the accuracy in comparison with existing techniques.
After that, a quantum mechanics inspired tracking method is proposed to address the abrupt motion problem. The proposed method uses some interesting phenomena in the quantum world, where the person is allowed to exist at multiple positions simultaneously. The results show a significant improvement in reducing the tracking error and in reducing the tracking delay.

Item Type:	Thesis (Doctoral)
Faculty:	School of Computing and Digital Technologies > Computing
Depositing User:	Library STORE team
Date Deposited:	11 Jun 2019 12:36
Last Modified:	11 Jun 2019 12:36
URI:	https://eprints.staffs.ac.uk/id/eprint/5697

Actions (login required)

View Item