A multiple description scalable coding scheme for adaptive video streaming over next generation cellular networks

Abstract

Nowadays, billions of users are connected to the mobile infrastructure offering a plethora of emerging services, and video streams are among the main contributors to this unprecedented traffic explosion. The paradigm of dynamic spectrum access (DSA) was coined and has become a widely accepted technology to boost the capacity of next generation networks. In this work, we propose an enhanced video coding implementation jointly with a DSA protocol wherein the sender actively picks up the most appropriate access strategy contingent on the instantaneous spectrum observations. The legacy frequencies pool has been extended to include the spectral ranges left unused in the television (TV) bands. A specific coding framework using a concatenation of multiple description coding and layered coding is investigated. The video is decomposed into two sub-streams through even and odd frames separation, then the resulting sub-streams are encoded into a scalable structure producing two independent and cumulative descriptions. The sender chooses to make use of the extended spectrum pool exclusively when the default frequencies are no more advantageous in terms of network capacity. Through extensive simulations, we compare the performance of the cellular network with and without the proposed DSA strategy. It has been proven that the proposed video transmission scheme using hybrid spectrum access provides a significant enhancement in network capacity up to 6.3Mbps and the capacity gain outweighs the disadvantage of interference incidents due to licensed-cognitive, co-channel and adjacent channel coexistence, besides retrieving an improved perceived video quality at the customer side illustrated by an increment of up to 6.8dB (0.176, respectively) in terms of the achieved PSNR-HVS-M (MS-SSIM, respectively).