Year: 2024
Type: dissertation
Abstract: The consistent growth of smartphones and the smart speaker market has played a vital role in establishing high-quality, far-field speech communication as a feasible alternative to traditional handsets. The prevalence of smartphones, multiple smart speakers and various computing platforms in many homes has led to much greater flexibility in how and where users communicate. This trend is expected to persist as next-generation voice and media services, such as Augmented Reality (AR)/Virtual Reality (VR), become more common. Additionally, key services such as financial and healthcare, for which flexible, high-quality communications are crucial components, are transitioning online. Despite these trends, comparatively little has been done to offer a converged communications experience once a speech call session is in progress. For example, the simple act of switching an ongoing call from a smartphone to a smart speaker is often a highly manual process. Pervasive communication systems aim to address this by providing a seamless, flexible communication experience across multiple devices and, where required, multiple networks. The primary contributions of this thesis provide solutions for both vertical handovers (transitions between heterogeneous networks) and horizontal handovers (transitions within a homogeneous network domain, specifically involving smart devices connected to the same network). This work uses a supervised machine learning based approach to predict user's transitions between the networks, and thus, overcome interruptions in speech due to signalling handover. For device handovers, this work proposes processing multivariate signalling features with time-series prediction algorithms and the deep learning techniques to accurately determine the most suitable device for the user for the handover. Additionally, this thesis considers how the Session Initiation Protocol (SIP) can be used in IP telephony systems where a seamless transition is required in a scope of handover between the networks and between the devices, while also proposing new solutions to achieve seamless session handovers.
Author: Mitnala, Vijaya Nirmala
Institution: University of Essex
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7043409016