Abstract
In recent years, there has been tremendous growth in digital multimedia technologies, from voice to data to video, and the recent, but growing, demand of supporting diverse quality of service (QoS) guarantees. It places new demands for future wireless networks in utilising the available radio resource in a more efficient and effective way. The key to this demand is the involvement of efficient radio resource management (RRM), to provide various QoS support for multimedia service delivery. Due to the unique broadcast nature and ubiquitous coverage of satellite communication system, the synergy between satellite networks and terrestrial networks provides new opportunities for delivering point-to-multipoint (or one-to-many) multimedia content to a large audience spread over extensive geographical area. It is expected that the satellite component will play a complementary, but essential, role in delivering multimedia data to those areas where the terrestrial high-bandwidth communication infrastructures are, either economically or technically, unreachable. The emphasis throughout this research is on the potential optimization techniques pertinent to the RRM to facilitate multimedia broadcast/multicast service (MBMS) delivery over the satellite digital multimedia broadcasting (SDMB) system, which has emerged as one of the most promising approaches for this mission. We concentrate on the algorithm development and performance evaluation of RRM strategies implemented at the radio access layer in the SDMB system, aimed at the efficient delivery of multimedia applications to mobile users at satisfactory QoS. Firstly, we have developed a novel two-level channel multiplexing scheme for the radio resource allocation (RRA), which is capable of optimally performing both logical and transport channel multiplexing via two new optimization algorithms, namely optimum estimation algorithm (OEA) and power-oriented adaptation (POA) algorithm. Simulation results prove that the proposed scheme can effectively improve the performance in terms of channel utilisation, power consumption as well as transmission capacity. Secondly, we have investigated the optimization of packet scheduling algorithms via the various adaptations of a proportional differentiation model. By taking into account multiple performance measures, e.g., buffer occupancy, queuing delay and data rate, several novel algorithms, i.e. buffer-length related queue (BLRQ), delay differentiation queue (DDQ) and combined delay and rate differentiation (CDRD), are introduced for performing the packet scheduling task in SDMB. Their performance has been evaluated via simulation means and compared with existing schemes. It is demonstrated that the proposed proportional differentiation packet scheduling schemes can achieve significant performance improvements in queuing delay, jitter and channel utilisation. Finally, we further optimize the packet scheduling schemes by using cross-layer design and adaptive optimization mechanisms. Cross-layer joint priority queue (CJPQ) and adaptive multidimensional QoS-based (AMQ) packet scheduling algorithms are introduced within this context and investigated via simulations. It is shown that these proposals can significantly improve the QoS performance amongst heterogeneous competing flows in terms of both scheduling efficiency and fairness, offering better flexibility and scalability features.