A distributed decision making method in cognitive radio networks for spectrum management

In order to make full utilization of the scarce spectrum resources for Cognitive radio networks, secondary users are expected to exploit the available spectrum of primary users. However, when there are several spectrum bands available, how to select an appropriate one for the secondary user according to the spectrum quality and the QoS requirements of different kinds of applications is a new challenge. In this paper, we propose a new Automatic distributed spectrum decision (ADSD) method to solve this problem. ADSD considers multiple spectrum characterization parameters, in particular, the primary users' arrival probability, to estimate the quality of the available spectrum bands. A weight auto-generation mechanism is included to automatically determine the weights of different parameters, thus avoiding the difficulty and irrationality when relying on the users to specify the weights directly. In addition, in conjunction with the reconfiguration mechanism, ADSD can reduce the rate of spectrum handoffs by reconfiguring the transmission parameters rather than making a new decision for the existing transmission. Simulation results show that without any users' interference, ADSD can automatically select the appropriate spectrum for transmission and significantly improve the Cognitive Radio network performance in terms of throughput and the spectrum handoff rate.