TR2015-132

Compressive Sensing for Loss-Resilient Hybrid Wireless Video Transmission


Abstract:

The quality of wireless video delivery is susceptible to wireless channel instability, including channel fading, interference, noise, and packet loss. To improve the video quality in such wireless channels, analog transmission schemes have been proposed recently. However, the existing analog schemes have two drawbacks in high energy of video signals and loss resilience. To overcome the issues, we propose a new hybrid digital-analog transmission scheme, which jointly uses digital coding and analog coding based on compressive sensing. Digital coding generates the residual signals between the original and the encoded signals to decrease the energy of data for analog coding. Compressive sensing redistributes the energy across whole video packets to increase the resistance to packet loss. We show that our proposed hybrid scheme improves video quality by up to 11.6 dB compared to the existing analog scheme in an erasure wireless channel. In addition, our proposed scheme also improves video quality by 1.5 dB compared to the existing hybrid scheme.