Cognitive Radio Network System Of Spectrum Allocation Study

The development of radio communication technology leads to the shortage of spectrum resource. However, recent researches show that the lack of wireless spectrum is a relative term, since a small amount of spectrum has been implemented quite a lot of services, while other parts of spectrum waste much, or say have fewer efficiency of utility. Point against such problem, the concept of cognitive radio has been proposed in order to improve spectrum efficiency and use spectrum in flexibility.The cognitive radio is an intelligent wireless radio which can sense the outside world and learn the information of surroundings by its artificial intelligence. It can communicate on a reliable basis at anytime and anyplace while using the frequency efficiency by changing real time operation parameters such as transmit power, carrier frequency to accommodate the internal to the outside world. Cognitive radio accesses the spectrum opportunisticly, while keeps the interference to the licensed user under a certain level. And how to allocate spectrum among cognitive radios is a research focus.Firstly, the dissertation introduces the basic concept and research of cognitive radio as the background and theory basis.Since the main content of this dissertation is dynamic spectrum allocation in cognitive radio system, thus in the second part, we introduce spectrum sharing model of cognitive radio at the beginning, as well as the allocation principle and technologies, respectively. Then, we give an explanation of two primary allocation models, wherein one is based on game theory and the other one on graph coloring theory.Besides the allocation models mentioned above, there are literatures also referring to some others, such as employing the theory of ant colony or shoal of fish into spectrum access and allocation process. The third part of this dissertation proposes a novel matching algorithm of spectrum allocation based on Stable Marriage Problem. According to simulation, this algorithm can converge to a deterministic point, a low computation complexity and a fast convergence speed can also be found. Furthermore, the proposed method can bring a higher satisfaction degree in the investigation of the performance of the whole cognitive radio networks.CSGC algorithm is a common graph coloring allocation method, which considers the difference in spectrum benefit and in spectrum interference. Furthermore it makes research in the comparison between collaborative and non-collaborative modes. However, this method has not considered allocation unfairness in the respect of different available spectrum number in each cognitive radio. Thus, the fourth part of this dissertation proposes a fairness algorithm combining color degree. According to computer simulation, the proposed method can have a higher fairness degree over the MSB algorithm in the condition that the two methods'total utilities should be near.