Temporal spectrum access in cognitive radio network

Dynamic spectrum access and spectrum sharing plays a critical role in cognitive radio networks. In this thesis, we apply two approaches for the spectrum allocation: game theory approach and learning automata approach. In the game theory approach, we propose a novel approach to a cooperative spectrum sharing game with varying number of cognitive radio pairs. Each radio pair is considered as a player in the game using the potential game model. Simulations show that by controlling the attendance of the new players in the game, the convergence speed of the game can be improved dramatically. In the learning automata approach, we apply the learning automata techniques to enable a cognitive radio to learn and make decision on channel selecting from a set of primary channels. We consider the situation where the availabilities of the channels are variable, thus the radio environment is non-stationary. We propose an adaptive learning algorithm that enables the cognitive radio to monitor the changes in the radio environment, thus always select the best available channel after a long run.Master of Engineering (SCE