Eldaw, Muawya Habib Sarnoub (2019) Analytics of human presence and movement behaviour within specific environments. Doctoral thesis, Birkbeck, University of London.
|
PDF
PhD_thesis_M_Eldaw_18Dec2019.pdf - Full Version Download (2MB) | Preview |
Abstract
The vast amounts of detailed information, generated by Wi-Fi and other mobile communication technologies, provide an invaluable opportunity to study different aspects of presence and movement behaviours of people within a given environment; for example, a university campus, an organisation office complex, or a city centre. Utilising such data, this thesis studies three main aspects of the human presence and movement behaviours: spatio-temporal movement (where and when do people move), user identification (how to uniquely identify people from their presence and movement historical records), and social grouping (how do people interact). Previous research works have predominantly studied two out of these three aspects, at most. Conversely, we investigate all three aspects in order to develop a coherent view of the human presence and movement behaviour within selected environments. More specifically, we create stochastic models for movement prediction and user identification. We also devise a set of clustering models for the detection of the social groups within a given environment. The thesis makes the following contributions: 1. Proposes a family of predictive models that allows for inference of locations though a collaborative mechanism which does not require the profiling of individual users. These prediction models utilise suffix trees as their core underlying data structure, where predictions about a specific individual are computed over an aggregate model incorporating the collective record of observed behaviours of multiple users. 2. Defines a mobility fingerprint as a profile constructed from the users historical mobility traces. The proposed method for constructing such a profile is a principled and scalable implementation of a variable length Markov model based on n-grams. 3. Proposes density-based clustering methods that discover social groups by analysing activity traces of mobile users as they move around, from one location to another, within an observed environment. We utilise two large collections of mobility traces: a GPS data set from Nokia and an Eduroam network log from Birkbeck, University of London, for the evaluation of the proposed models reported herein.
Metadata
| Item Type: | Thesis |
|---|---|
| Copyright Holders: | The copyright of this thesis rests with the author, who asserts his/her right to be known as such according to the Copyright Designs and Patents Act 1988. No dealing with the thesis contrary to the copyright or moral rights of the author is permitted. |
| Depositing User: | Acquisitions And Metadata |
| Date Deposited: | 28 Jan 2020 15:04 |
| Last Modified: | 06 Mar 2025 06:47 |
| URI: | https://eprints.bbk.ac.uk/id/eprint/40458 |
| DOI: | https://doi.org/10.18743/PUB.00040458 |
Statistics
Additional statistics are available via IRStats2.
Archive Staff Only (login required)
