An adaptive pursuit strategy for dynamic spectrum assignment in narrowband PLC

摘要

In this paper the process of spectrum assignment in narrowband power line communication is studied. Due to the frequency-selective and time-variant nature of the PLC channel, spectrum assignment must be performed dynamically throughout the transmission. This requires constant channel state information at the PLC transmitter which in turn increases undesired overhead in the system and degrades throughput. An adaptive pursuit strategy has been proposed in order to overcome this problem by enabling the transmitter through a reinforcement learning method to select the best part of the spectrum dynamically and without prior knowledge of the channel. Simulation results show a fast adaptation to the channel conditions which results in an increase in spectral efficiency and improves the performance of the transmission.