Knowledge-Enhanced Mobile Video Broadcasting Framework With Cloud Support

摘要

The convergence of mobile communications and cloud computing facilitates the cross-layer network design and content-assisted communication. Mobile video broadcasting can benefit from this trend by utilizing joint source-channel coding and strong information correlation in clouds. In this paper, a knowledge-enhanced mobile video broadcasting (KMV-Cast) is proposed. The KMV-Cast is built on a linear video transmission instead of a traditional digital video system, and exploits the hierarchical Bayesian model to integrate the correlated information into the video reconstruction at the receiver. The correlated information is distilled to obtain its intrinsic features, and the Bayesian estimation algorithm is used to maximize the video quality. The KMV-Cast system consists of both likelihood broadcasting and prior knowledge broadcasting. The simulation results show that the proposed KMV-Cast scheme outperforms the typical linear video transmission scheme called Softcast, and achieves 8 dB more of the peak signal-to-noise ratio (PSNR) gain at low-SNR channels (i.e., -10 dB), and 5 dB more of PSNR gain at high-SNR channels (i.e., 25 dB). Compared with the traditional digital video system, the proposed scheme has 7 dB more of PSNR gain than the JPEG2000 + 802.11a scheme at a 10-dB channel SNR.