Error-Resilient Video Multicast with Layered Hybrid FEC/ARQ over Broadband Wireless Networks

Abstract

Video multicast over broadband wireless networks suffers from packet losses induced by fading wireless channels and user heterogeneity in channel conditions within a multicast group. A promising solution to these problems is the use of layered hybrid FEC/ARQ for scalable video multicast. However, how to allocate the radio resources to multiple video layers and how to address the cross-layer combination of application layer hybrid FEC/ARQ and physical layer MCS (modulation and coding schemes) for each video layer, is not a trivial issue. We prove that this problem is NP-hard and propose an optimal Error-resilient Video Multicast (ERVM) framework in infrastructure-based broadband wireless networks. To avoid user heterogeneity and feedback implosion, we use a weighted designated user group to send light-weight feedback messages. The ERVM algorithm is based on the dynamic programming method with a pseudo-polynomial run time, and can get the optimal transmission configuration to maximize the expected utility for the designated user group, which is a probabilistic function of the received video rate. Simulation results show that our ERVM algorithm offers significant improvements over the conventional single-layer FEC/ARQ scheme and the layered FEC scheme. Furthermore, our weighted designated user group with light-weight and accurate feedbacks is better than other user feedback schemes.