A Near Optimal QoE-Driven Power Allocation Scheme for Scalable Video Transmissions Over MIMO Systems

The rapid increasing demands of wireless multimedia applications have boosted the developments of video delivery technologies with cross-layer designs, driven by optimizing quality of experiences (QoEs) of end users. In this paper, a near optimal power allocation scheme, targeting at maximizing QoE, is proposed for transmitting scalable video coding (SVC) based videos over multi-input multi-output (MIMO) systems. Both transmission errors in the physical (PHY) layer and video source coding characteristics in the application (APP) layer are jointly considered in the proposed scheme. A near optimal solution is achieved by decomposing the original optimization problem into several convex optimization sub-problems. Detailed algorithms with corresponding theoretical reasoning are provided. Since forward error corrections (FEC) techniques are widely implemented in modern wireless communication systems, the proposed scheme is further extended to the systems with Reed-Solomon (RS) code and a more practical approach with different modulation and coding schemes (MCSs). The near optimality of our proposed scheme, in terms of measured utilities, is shown by comparing with the exhaustive searched optimal solutions. Simulations with real H.264 SVC video traces demonstrate the effectiveness of our proposed scheme by comparing with other existing schemes in terms of well-accepted video quality assessment methods, such as peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) index.