Adaptive Hybrid Digital-Analog Video Transmission In Wireless Mobile Networks

In wireless mobile networks, channel fluctuations caused by multipath propagation, shadow fading and Doppler frequency shift will result in severe fluctuations in the received video quality, namely the cliff effect. Current analog video transmission schemes can eliminate the annoy cliff effect. However, there still exist two limitations in analog video transmission:1) the low coding and transmission efficiency;2) unable to work effectively in fading channel and cope with the adverse effect caused by channel fluctuation. To cope with the aforementioned concerns, we develop A-HDAVT for wireless video streaming applications in this research. A-HDAVT attempts to achieve excellent rate-distortion-capacity performance for digital transmission as well as graceful degradation performance for analog transmission. In A-HDAVT, a joint power allocation scheme is designed to achieve simultaneously reliable digital transmission for reference frames and adaptive analog transmission for High-pass frames based on channel prediction. Effectively, the proposed A-HDAVT is to exploit both content diversity and channel diversity in order to optimize power allocation during video transmission. The main contributions of this research are:1) In A-HDAVT, a power allocation strategy for hybrid digital-analog video transmission is developed. The critical Low-pass frame is allocated with sufficient power and transmitted in reliable digital transmission to ensure basic video quality even in poor channel states. The remaining power is then allocated to the High-pass frames in analog transmission to achieve graceful degradation performance. This enables robust video transmission with guaranteed minimum acceptable video quality as well as maximum possible video quality under wireless fading channels.2) In analog transmission, an algorithm for channel Prediction based Adaptive Power Distortion Optimization (P-APDO) in A-HDAVT is designed to combat channel fading and to improve the user perceived video quality. With P-APDO, each High-pass frame is hierarchically decomposed and proper power are allocated to various data units according to their relative importance. This enables the exploitation of video content diversity at multi-resolution level.3) In multi-user scenarios, an extended P-APDO is also developed assuming that each user is transmitting with an independent sub-channel. In this case, the extended P-APDO is able to exploit both content diversity and channel diversity among users. Both power pre-allocation and power re-allocation are executed. Power pre-allocation is performed to allocate power among video data of all users based only on their content importance. However, power re-allocation is applied to re-allocate power among video data being transmitted with additional consideration of the predicted channel states for each particular user.4) Various simulations are carried out to evaluate the performance of the proposed A-HDAVT under various channel fading conditions in both single user and multi-user scenarios. The performance of A-HDAVT is also compared with that of SoftCast, ParCast under the same channel condition in both single-user and multi-user scenarios to demonstrate the inherent advantage of A-HDAVT when the wireless channels are experiencing the inevitable fading in all practical video transmission implementations.