Adaptive Source-FEC Coding for Energy-Efficient Surveillance Video over Wireless Networks

Jiyan Wu,Bo Cheng,Ming Wang

DOI: https://doi.org/10.1109/tcomm.2017.2785252

IF: 6.166

2018-01-01

IEEE Transactions on Communications

Abstract:Video surveillance has become an important Internet multimedia application to provide live streaming services. Energy efficiency is critical to guarantee the service time and quality of power-intensive video streaming on wireless surveillance systems. In particular, video coding and data communication constitute the majority of power dissipation in embedded multimedia devices driven by capacity-limited batteries. However, the complex power characteristics and time-varying channel status pose crucial challenges on enabling low-power high-quality video streaming. To address these critical problems, this paper presents a power-efficient and network-adaptive (PENA) framework for source-forward error correction (FEC) coding in embedded systems. First, an analytical framework is developed to characterize the rate-distortion-power tradeoff for video encoding and data transfer. Second, a joint rate control and FEC coding solution is proposed to maximize video quality under power constraint. Distinct from the existing source-FEC coding algorithms, PENA is able to effectively leverage the video rate-distortion model and system power features. We conduct the system implementation and performance evaluation with real embedded surveillance devices over wireless networks. Experimental results demonstrate PENA achieves appreciable improvements over the reference source-FEC coding schemes in terms of power conservation, perceived video quality, and end-to-end delay.