Research On Spectrum Allocation And QoS Guarantee Mechanisms Over Cognitive Radio Networks

With the development of new technologies(wireless sensor networks,Internet of things,etc.)and the popularization of intelligent mobile terminals in recent years,the traffic in wireless communications has increased rapidly.Therefore,the spectrum shortage problem has become increasingly serious.Cognitive Radio Networks(CRNs)which combine the advantages of cognitive radio technology and Ad Hoc networks,have been viewed as an intelligent network models and have self-adaption abilities to the complex network environment in applications.In CRNs,mobile terminals could obtain ample wireless bandwidth in transmission through resource management and dynamic spectrum access techniques,which could effectively optimize the end-to-end performance and resource utilization.After sensing and understanding the surrounding environment,the Secondary Users(SUs)could carry on their flow transmission according to the Quality of Service(QoS)requirements at the premise of controlling the interference to Primary Users(PUs).Due to features of CRNs,there are some challenges in SUs' QoS guarantee,and the challenges are mainly including the following aspects: 1)Considering the diversification of different users' requirements,how to effectively allocate the resources to coexisted users in order to improve spectrum utilization.2)Compared with PUs,SUs have lower priority at the access to spectrum resource,how to ensure their transmission quality without affecting PUs' transmission.3)How to design stable and efficient routing protocols considering the problems such as the time-varying feature of available spectrums,the dynamic change in network topology,the link instability and the divergence of users' QoS requirement,etc.To overcome these above-mentioned challenges,the research in this thesis starts from the influence of spectrum management strategy to the network performance optimization in CRNs.And the research includes: spectrum allocation strategy based on bidding game,fairness-based resource allocation strategy for SUs with diversified QoS requirements,auction based spectrum allocation strategy,multi-metric cross-layer routing establishment and maintenance,cluster-based routing algorithm for video transmission.Our objective is to realize that the resource allocation strategy could dynamically adjust with environment,thus we could optimize SUs' flow transmission efficiency,and users could obtain the required QoS guarantee.The work and innovation of this thesis mainly include the following aspects:Considering the heterogeneous wireless services of cognitive radio networks,this thesis proposes a weighted interference graph and bidding game based spectrum allocation model,and the allocation strategy under dynamic environment.We analyze the features of CRNs and propose a unified resource allocation framework connected with QoS requirements.Primary networks provide several channels with different parameters,and SUs could choose the spectrum according to QoS requirements.Considering the competition between SUs in spectrum allocation,this thesis designs a weighted interference graph based grouping algorithm to divide SUs into several independent groups according to resource requirement and construct the non-interference user sets.We propose the bidding game based spectrum allocation mechanism,both buyer's and seller's revenues are analyzed and accordingly the optimal allocation scheme is derived.Managers could control the supply-and-demand relation of spectrums through adjusting the unit bandwidth price,thus achieving the balance of the spectrum resources market.We prove the existence of Nash equilibrium and deduce that the strategy could achieve equilibrium pricing under competitive environment,also we analyze the economic attributes(users' individual rationality and trustfulness)of spectrum allocation.Simulation and analysis show that,the allocation model proposed could optimize the system revenue and guaranteeing users' profit,provides a feasible solution to improve the spectrum utilization and resource management efficiency.This thesis proposes a classification mechanism for users according to QoS requirements,and proposes a fairness based resource allocation strategy to guarantee that different users could obtain the required transmit quality in CRNs.We formulate the resource allocation problem as a constrained optimization problem with the objection of maximizing all SUs' sum transmit rate.We provide the system utility function with power and channel constraints.Considering the high complexity of spectrum allocation problem,we transform the optimization function into two sub-problems,and find the suboptimal alternative to decrease the complexity.In order to guarantee the allocation fairness between different types of users,we design a resource allocation strategy that could adjust the fairness degree dynamicly through penalty functions.Also we propose the fairness base resource allocation algorithm towards diversified user requirements,the proposed algorithm could achieve efficient resource allocation between different users thus motivate users' participation.A QoS layering based spectrum allocation algorithm is proposed,each SU's channel selection is independent of others,and the purpose of spectrum allocation is to maximize the system throughput.The service classification related with network parameters is realized,and users could select spectrums according to QoS levels,therefore the problem of competition between users caused by requirement divergence could be solved.When the manager receives the access request,it chooses the winners through an auction-based algorithm and calculates how to maximize the system revenue with the limited spectrum resource.This strategy could provide QoS guarantee for users and ensure the rationality of spectrum allocation.In dynamic changing environment,the time-varying feature of spectrum availability and users' mobility will cause link instability problem between different nodes,thus affecting the stability of multi-hop routing.This thesis discusses the multi-hop cognitive routing protocol,and designs the routing strategy combined with spectrum opportunity for cognitive ad hoc networks.This thesis points out the multi-hop cognitive routing problem through network topology,and proposes a multi-metrics optimization strategy combined with users' special requirements.We analyze the Pareto optimal solution to the muti-object optimization problem,thus we could find the optimal routing with minimum overhead.The weights of routing metrics are defined according to the feature of traffic flow,and a cross-layer optimization strategy of channel and route switch is proposed.With the analysis of performance parameters,a multi-metric routing protocol with routing recovery mechanism is proposed.The path with minimum routing overhead is viewed as the target path in order to optimize the end to end network performance.Considering the feature of video streaming(delay-sensitive and high bandwidth,etc.),a clustering based routing algorithm for multimedia cognitive networks is proposed.In order to ensure the cooperation efficiency of multiple nodes in CRNs,we propose an importance based cluster head election rule on the basis of the clustering model in cognitive radio Ad Hoc networks.Combined with the dynamic spectrum allocation characteristics,the thesis analyzes the clustering scale of all nodes to minimum the transmission distortion.We also propose the cognitive clustering based routing algorithm,and validate the effectiveness of the proposed algorithm through simualtion.