The Research On Key Technologies Of Spectrum Sense And Access Schemes For Wireless Cognitive Networks

Wireless network has been developed tremendously in the last decade, which leads to the situation that various wireless services have exceeded the corresponding wired services. As the traditional static spectrum usage policy can't support the huge demand of wireless service, a new open spectrum access policy has been proposed. The open spectrum strategy allows unlicensed users to access the spectrum opportunistically if they do not interfere with the licensed users. The cognitive radio is the key device to achieve the dynamic spectrum access goal. The cognitive network which mainly consists of cognitive radio users has been widely regarded as the main network struction for the next generation of wireless network.Differing from traditional wireless networks, spectrum sensing and access are two fundamental processes for resource management. Spectrum sensing is the process of discovering available spectrum while spectrum access is the process of using the spectrum. Spectrum sensing is the fundamental process of cognitive network's resource management and spectrum access is essential to the spectrum utility. As the interference is determined by the interference temperature, the unlicensed user must control their transmitting power to guarantee the interference limit of licensed user. The resource management mainly consists of how to design the sensing,access and power adjustment schemes. These new schemes are different from traditional solutions and they need to be redefined and researched.This dissertation focuses on the critical problems of spectrum management in cognitive networks. These problems are: optimal balance of spectrum sensing and access for homogeneous cognitive networks; optimal spectrum access for heterogeneous cognitive networks; cognitive networks'capacity estimation; power control and spectrum allocation for interference temperature based cognitive networks. For each problem, the detailed complexity analysis is described and proved to be incomputable within polynomial time. Then approaximated algorithms are designed to solve these problems efficiently. The goals of all algorithms are to improve the spectrum utility by minimum computation complexity.For the optimal balance of spectrum sensing and access problem, we have established a sensing and access model on the assumption that sensing time is not neglectable. As the sensing time will determine the spectrum utility, the sensing efficiency is critical. Due to the contention relation among traffic links, the links which are within each other's interference range can not access the spectrum simultaneously. It is essential to the contending links to sense heterogeous spectrum without conflict. A new clique partion scheme is firstly proposed to solve the sensing scheduling problem. By optimally allocating spectrum between construed cliques and within cliques, the sensing and access balance is solved efficiently. Through detailed simulation, the results have demonstrated the performance of algorithm.Considering a cognitive network which licensed users have heterogeous spectrum availability, how to design the spectrum access scheme is critical to the spectrum utility. We proposed two access schemes on the foundation of maximum weighted matching for general graph and maximum weighted independent set theory. Various simulation results show that the spectrum utility is greatly improved.The network capacity analysis is always one of the most important theoretical aspects of wireless networks. Upper and lower bound is the target for these researches. As the highly dynamical characteristic of cognitive network environment, the general bound value is insufficient for the reference of network management. For this, we have proposed a clique based capacity estimation scheme. Through the inner capacity estimation of cliques distributedly, the near-optimal capacity estimation value is acquired. The capacity estimation value is closely related to the optimal value at most network simulation scenarios. The results demonstrated the algorithm is stable and effective.As the interference protocol model can't meet the demand of interference temperature requirement, we have proposed two schemes to solve the interference limit problem by jointly considering power control and spectrum allocation. One is a heuristic algorithm which considers a new parameter of relative interference distance; another is an iterative algorithm which can adjust the power value more precisely to the optimal scheme. The two schemes can guarantee the SINR requirement and improve the spectrum utility efficiently. Especially the iterative algorithm is a fast converging, stable and effective solution.In summary, this dissertation focuses on spectrum management and systematically investigates key problems which fundamentally impact the spectrum utility. Through detailed theoretical and experimental analysis of four essential problems, reliable and efficient solutions are proposed respectively. These solutions can help cognitive users perform the sensing, access and power adjustment process under different network scenarios. The research results can provide a basis for the research of next multi-hop and multi-service cognitive networks. We believe these results will also enhance the protocol design process of wireless cognitive network.