Video Transmission over LTE-Unlicensed

Abstract

The witnessed proliferation in the demand on wireless spectrum over the past years is anticipated to lead to a bandwidth scarcity problem. The situation is further aggravated by the combined increase in the demand on the delivery of high definition multimedia content over wireless networks. Recently, it was proposed to extend the LTE operation into the unlicensed spectrum (LTE-U) in an attempt to mitigate the expected scarcity problem. However, extending the operation of LTE to the unlicensed spectrum is expected to degrade the performance of other incumbent technologies, especially WiFi. While several LTE-U channel access mechanisms for enabling the coexistence of LTE with WiFi, in the unlicensed spectrum, have been proposed and evaluated, very few addressed the problem of video streaming over LTE-U. Therefore, this thesis further explores the design of efficient LTE-U channel access mechanisms from the perspective of video streaming. Specifically, we propose two channel selection schemes for video transmission over LTE-U. The first is a static channel selection scheme in which the LTE-U user equipment (UEs) are assigned to channels based on their frame type. The second one is an adaptive channel selection scheme in which UE frames are assigned to channels based on the solution to a multi-objective optimization problem that takes into account the UE's Quality of Experience (QoE) requirements while striving to minimize the interference caused to WiFi. Simulation results show that, compared to a scheme that employs an effective channel access mechanism but assigns frames to channels randomly, our proposed QoE-based channel selection scheme achieves a considerably better performance, in terms of improving the received video quality and reducing the inter-frame delay for both LTE-U and WiFi users. These results confirm the conclusion that while extending the operation of LTE-U to the unlicensed spectrum does indeed come with great benefits for LTE networks, that extension should be accompanied by smart channel selection schemes that take both the nature of the UEs data and the channel conditions into account.