Joint Power and Bandwidth Allocation for 3D Video SoftCast

Abstract

Unlike digital video transmission systems, SoftCast avoids the cliff effect and achieves a linear video quality transition that is commensurate with the wireless channel conditions. When SoftCast is applied to three-Dimension Video (3DV) transmission, resource allocation issues arise: 1) allocating the limited power budget to texture and depth to achieve the optimal overall quality. 2) distributing the suitable number of texture and depth chunks to adapt to bandwidth constraints. This work aims to efficiently solve the optimal joint power and bandwidth allocation problem. First, a power-distortion optimization problem is formulated to calculate the optimal Power Allocation Ratio (PAR) between texture and depth, then mapped to an unconstrained problem and solved using the Lagrangian multiplier. Finally, based on the closed-form of the optimal solution, an iterative algorithm is proposed to choose the suitable number of texture/depth chunks for a given bandwidth constraint. The proposed method achieves better performance than its counterpart default fixed-ratio power allocation between texture/depth. Further, we observe a graceful video quality transition with the improvement of channel conditions under bandwidth constraints.