Predictive channel estimation for optimized resources allocation in DVB-S2 networks

Exploiting the Adaptive Coded Modulation (ACM) mechanism leads to a more spectrum-efficiency since the transmission parameters are dynamically adjusted to accommodate the channel conditions. This improvement is, however, limited when used in satellite environments due to the time-varying nature of the channel conditions. This indeed make harder for the transmitter to match instantaneously the current network conditions with the appropriate ACM parameters. The high delays entailed by accurate network condition measurement and end-to-end reporting, and the feedback propagation are a major hurdles to overcome. In this paper we propose to use channel prediction, instead of using the instantaneous channel state feedback, as a mean to combat the counterproductive effects caused by the feedback latency. Based on the predicted values, we propose a new modulation and code rates (MODCOD) selection algorithm, which considers the measurement impairment by introducing some margins allowing the selection of more robust MODCOD. Further, we propose a sliding window technique as a mean to combat the detrimental effects of the frequent MODCOD switching and the entailed oscillations in the system performance. Simulation results elucidate that the proposed scheme allows a more efficient network resources´ utilization while maintains the Bit Error Rate (BER) within an acceptable level.