Bandwidth Estimation and Optimisation in Rain Faded Dvb-rcs Networks

Broadband satellite communication networks operating at Ka band (20-30 GHz) play a very important role in today’s worldwide telecommunication infrastructure. The problem, however, is that rain can be the most dominant impairment factor for radio propagation in these frequency bands. Allocating frequency bandwidth based on the worst-case rain fading leads to the waste of the frequency spectrum due to over reservation, as actual rain levels may vary. Therefore, it is essential that satellite systems include adaptive radio resource allocation combined with fade mitigation techniques to efficiently counteract rain impairments in real-time. This thesis studies radio resource management problem for rain faded Digital Video Broadcast-Return Channel via Satellite (DVB-RCS) networks. This research stems from taking into account two aspects in the bandwidth estimation and allocation process: the consideration of multiple rain fading levels; and the geographical area size where users are distributed. The thesis investigates how using multiple rain fading levels in time slot allocation can improve bandwidth utilisation in DVB-RCS return links. The thesis presents a mathematical model to calculate the bandwidth on demand. The radio resource allocation is formulated as an optimisation problem, and a novel algorithm for dynamic carrier bandwidth and time slots allocation is proposed, which works with constant bit rate type of traffic. The research provides theoretical analysis for the time slot allocation problem and shows that the proposed algorithm achieves optimal results. This thesis also studies Return Channel Satellite Terminals (RCSTs) geographical distribution effects on bandwidth demand and presents a novel mathematical model to estimate the maximum instantaneous bandwidth demand for RCSTs randomly distributed over a geographical area in a satellite spot beam. All the proposed algorithms have been evaluated using a novel simulation with historical rain data.