Multimedia streaming over wireless channels

Official URL: http://risc01.sabanciuniv.edu/record=b1147599 (Table of Contents)

The improvements in mobile communication systems have accelerated the development of new multimedia streaming techniques to increase the quality of streaming data over time varying wireless channels. In order to increase multimedia quality, error control schemes are indispensable due to time-varying and erroneous nature of the channel. However, relatively low channel capacity of wireless channels, and dependency structure in multimedia limit the eectiveness of existing error control schemes and require more sophisticated techniques to provide quality improvement on the streaming data. In this thesis, we propose sender driven multimedia streaming algorithms that incorporate error control schemes of FEC, ARQ, and packet scheduling by considering media and channel parameters such as packet importance, packet dependencies, decoding deadlines, channel state information, and channel capacity. Initially, we have proposed a multi-rate distortion optimization framework so as to jointly optimize FEC rate and packet selection by minimizing end-to-end distortion to satisfy a specified Quality of Service under channel capacity constraint. Minimization of end-to-end distortion causes computational complexity in the rate distortion optimization framework due to dependency in encoded multimedia. Therefore, we propose multimedia streaming algorithms that select packet and FEC rate with reduced computational complexity and high quality as compared with multi-rate distortion optimization framework. Additionally, protocol stack of a UMTS cellular network system with W-CDMA air interface is presented in order to clarify the relation between proposed multimedia streaming algorithms and UMTS system that is used in simulations. Finally, proposed algorithms are simulated and results demonstrate that proposed algorithms improve multimedia quality significantly as compared to existing methods.