Predictive Location-Aware Network Selection and Handover in Heterogeneous 4G Networks: A Media-Independent Approach to Seamless Mobility
No Thumbnail Available
URL
Journal Title
Journal ISSN
Volume Title
School of Science |
Master's thesis
Checking the digitized thesis and permission for publishing
Instructions for the author
Instructions for the author
Authors
Date
2010
Major/Subject
Informaatiotekniikka
Mcode
T-115
Degree programme
Language
en
Pages
xxvi + 175
Series
Abstract
In this thesis we describe a method for predicting user behavior based on local movement patterns within a finite geographical area. This predictive location-aware method may then be utilized by the PLANS algorithm to enable optimal network selection (NS) according to an always best connected (ABC) paradigm, as well as by the PLAVHO algorithm to enable seamless vertical handover (VHO) between different radio access technologies (RATs). The work also describes the proof-of-concept prototype that was used for testing these methods, and presents the preliminary test results. Both ABC and seamless VHO play a key role in forthcoming fourth-generation (4G) all-IP networks (AIPNs), in which all services, even time-critical real-time applications, such as telephony, are Internet-based. Typically a 4G AIPN may comprise several heterogeneous RATs, such as hotspots of Wi-Fi wireless local-area networks (WLANs) placed within a ubiquitous UMTS or LTE-based wireless wide-area network (WWAN), with the hotspots generally offering improved service, e.g. higher bandwidth, or lower costs for the end-user, but only within their geographically limited coverage areas. Consequently, NS, especially the ABC paradigm, as well as VHO have been studied thoroughly during the last years, although the effects of movement and statistical movement patterns have, surprisingly, largely been neglected within the context of heterogeneous 4G networks, despite these having been hot research topics within previous-generation homogeneous public land mobile networks (PLMNs). Likewise, neither have energy consumption and seamless mobility received the attention they deserve, despite both being identified as critical aspects in utilization of mobile AIPN services. Therefore, using a Linux-based prototype, we examine the use of statistical movement patterns to predict user behavior, optimizing NS, while maintaining seamless operation through proactive make-before-break (MBB) VHO. The interfaces in this prototype implemented in Python, using Wi-Fi and 3G for connectivity, MIPv4 for maintaining active sessions, and GPS for positioning, are primarily based on the IEEE 802.21 standard for media-independent handover (MIH), thus facilitating potential implementation of this architecture in true 4G network environments. Finally, based on the preliminary test results collected using the prototype, predictive locationawareness appears to benefit both NS and VHO, potentially providing both greater mean bandwidth, and lesser mean latency, possibly also providing means for improving energy efficiency.Description
Supervisor
Simula, OlliThesis advisor
Weckström, TomKeywords
PLANS, rörlighet, PLAVHO, platsmedvetenhet, location-awareness, spårning, always best connected ABC, alltid bäst uppkopplad, seamless handover, medieoberoende handover, vertical handover, vektorkvantisering, vector quantization, klassificering, kd-tree, lokalisering, clustering, nearest neighbor, 3G, 4G, positioning, location service, GPS, Mobile IP, MIPv4, Linux, Python, Twister, SciPy, NumPy