Scalable Transmission Scheme Research For Wireless Video Multicast

With the maturity of4G communication technologies, the data transfer rate of wireless communication is rising significantly, and the storage capacity and processing capability of mobile devices and the popularity of smart devices are increasing. Traditional telecommunication services including message service and voice service have been unable to meet the diverse needs of users. Wireless video services over mobile devices are getting more and more popular.Wireless video services usually occupy large bandwidth resource and have strict restrictions on the wireless transmission delay. When multiple users watch the same video simultaneously, wireless video multicast is an efficient transmission scheme of video data, which can greatly save bandwidth resource of wireless communication system. However, the channel condition of each user is inconsistent, so how to provide the video quality adaptive to the channel condition of each user is a major challenge for wireless video multicast.According to the processing and transmission requirements of wireless video multicast, a scalable transmission scheme based on motion compensated temporal filter and two-dimensional discrete wavelet transform is proposed in this thesis. To meet the demands of scalability and robustness, the proposed scheme performs real-value representation, transform and transmission of the video stream, for the real-valued property can eliminate cliff effect and increase the resilience to channel noise.The technologies adopted in the proposed scheme include motion compensated temporal filter, two-dimensional discrete wavelet transform, Hadamard transform,64K quadrature amplitude modulation, run length coding and so on. Motion compensated temporal filter removes the inter-frame correlation and its motion alignment can reduce the temporal redundancy efficiently. Two-dimensional discrete wavelet transform removes the intra-frame correlation. Hadamard transform redistributes the signal energy to guarantee the same importance of all the packets. The dense constellation mapping of64K quadrature amplitude modulation is utilized to maintain the real-value property of video transmission signal. Moreover, metadata is extracted from the previous processes and transmitted in the conventional mechanism to facilitate reliable decoding. Motion vectors are further compressed using run length coding to reduce the metadata’s transmission overhead on the basis of ensuring accuracy.The proposed scheme and other typical transmission schemes have been simulated for performance comparison. Simulation results verify the scalability and robustness of the proposed scheme that it can avoid cliff effect and is resilient to channel noise and propagation loss.