Optimal power allocation auction for H.264/SVC coded wireless video transmission

Abstract

In this paper, we propose optimal power allocation schemes for quality maximization of the transmitted video streams in wireless multicast communication scenarios. For this purpose we employ parametric scalable video models, which model the rate and quality of the scalable streams as a function of the quantization parameter and frame rate. These are derived from the standard JSVM reference codec for the H.264 SVC/AVC and are hence readily applicable in practical scenarios. These models are subsequently employed to present a novel revenue maximization scheme in Orthogonal Frequency Division for Multiple Access (OFDMA) based wireless broadband 4G systems employing transmitted video steam quality based auction bidding models. The framework for optimal power allocation is formulated as a constrained convex optimization problem towards sum video utility maximization. We observe that as the demand for a video stream increases in broadcast/ multicast scenarios, higher power is allocated to the corresponding video stream leading to a gain in the overall revenue/ utility. Simulations illustrate that the proposed optimal power allocation schemes result in a significant performance improvement over the suboptimal equal power allocation schemes for scalable video transmission.