Mobile Edge Resource optimization for Multiplayer Interactive Virtual Reality Game

摘要

Edge computing has been regarded as an efficient approach to achieve the multiplayer interactive virtual reality (VR) game over wireless networks, where the game scenes can be rendered at the edge computing server and then the real-time video frames can be transmitted to the players. In order to ensure the fairness of the interacting players, we allocate the edge computing and wireless bandwidth resources for minimizing the average inter-player delay among different players. The proposed programming is under the constraints of the absolute delay requirements, the maximum frame per second (FPS) demands, the total bandwidth and rendering resources limits. The influence of prediction and pre-rendering field of views (FOVs) is also considered in the optimization problem to model the interaction among players. To tackle the non-convex problem efficiently, a sub-optimal algorithm which convert the original problem into several convex subproblems to optimize iteratively is designed. Finally, numerical results verify the proposed algorithm can reduce the average interplayer delay significantly with lower complexity, and it also reveals the impact of the content sizes and the channel state conditions on the edge resource allocation.