An adaptive cross layer resource allocation scheme for correlated wireless video sources

Abstract

In this work we study adaptive resource allocation for uplink transmission of correlated video sources. We consider a framework where multiple wireless sources transmit correlated information via a common base station. We introduce an optimization problem for sources to maximize the weighted sum of received quality of all videos, when source correlation can be used at the decoder in case of missing data. Each source finds its respective Multiple Access (MAC) parameters and performs packet selection. This is done with minimal information exchange with the base station. We model the quality of a decoded video as a piecewise linear function of qualities of the most correlated views, and verify the validity of the piecewise linear quality model for a two source case. We use this model to simplify the resource allocation method. We then compare performance of our correlated resource allocation to optimal resource allocation for independent sources, as well as to a baseline method. The simulations show that our proposed method results in higher average Y-PSNR than the optimal independent resource allocation in most channel conditions, without significative complexity increase in the base station or the source nodes.