Cooperative Spectrum Sensing Algorithm Based on Evolutionary Game Theory

Abstract

To address the inert behavior of cognitive users who do not sense to increase their throughput in the collaborative spectrum sensing process, a spectrum sensing model with a contribution reward and punishment mechanism is proposed to increase the motivation of spectrum sensing. The whole spectrum sensing process is normalized to the number of transmitted bits, and the energy consumption in the sensing phase is considered comprehensively and a reward and punishment function with a reward and punishment strength of 0.25 is introduced in the transmission phase to create an evolutionary game model for spectrum sensing, further select an evolutionarily stable strategy, and use a distributed algorithm to deal with the convergence problem for a given initial participation rate. The simulation results show that the stability of Dual-User cooperative spectrum sensing is the game strategy obtained by the fixed small penalty factor (perception, perception), and the game strategy obtained by the fixed large penalty factor (perception, nonperception) and (nonperception, perception). The superiority of the present reward mechanism is verified in the multi-user case, which can significantly improve the spectrum-aware participation rate and increase spectrum-aware motivation.