| （英） |
With the development of wireless communication technology, spectrum sharing needs due to the increase in the number of systems and the number of terminals. In a shared-wireless resource environment, interference avoidance is one of important technologies. Since the possibility of hidden terminals increases, it is expected to become difficult to avoid interference by simply using carrier sense, as was originally the case with CSMA-CA. Therefore, this study discusses a method for users to use spectrum resources in an autonomous and cooperative manner. In order to share wireless resources without terminals exchanging information with each other, the proposed method uses a multi-armed bandit algorithm, which is a type of reinforcement learning, to perform channel selection. Channel selection takes into account the residual resource length after transmission for immediate reward and the delay between packet generation and transmission when the channel is idle, and selects the channel with the smallest delay when a frequency resource matching its packet length is available. Simulation result shows that the proposed method improves the transmission success rate by reducing the possibility of other terminals' transmission opportunities being reduced by consuming part of the free resources that are excessively long compared to the length of the packets to be transmitted. |