On The Topology Control In Cognitive Radio Networks

In cognitive radio networks(CRNs),the spectrum efficiency can be improved via spectrum reutilization.Particularly,the cognitive users can access the spectrum if they do not interfere with the authorized users.However,due to the higher priority of authorized users,the transmission interruptions may occur between cognitive users when authorized users reclaim their channels.Such unpredictable activities may partition the network,which further results in intolerable packets delay and heavy collisions.Therefore,how to maintain an end-to-end transmission in CRNs considering the activities of authorized users? In the academia,the above problem is deemed as the connectivity guarantee problem of graph theory;and topology control is adopted to construct fault-tolerated network topologies.There exist two paradigms of CRNs: overlay and underlay.The overlay paradigm imposes SUs to access parts of the licensed spectrum unutilized by PUs.With the underlay paradigm,SUs are allowed to transmit with PUs on the same spectrum whenever they do not cause unacceptable interference to PUs.Existing research has only considered the overlay paradigm and guaranteeing the connectivity of CRNs when a channel is reclaimed.These limitations may result a lower spectrum efficiency and the network partition whenever multiple channels are occupied.To this end,this dissertation considers the underlay paradigm,and devotes to designing the proper topology control algorithms to maintain the connectivity of CRNs.The main contributions of the dissertation are summarized as follows:1.We reveal the channel-connected property of underlay CRNs.With considering the underlay paradigm,where the authorized and cognitive users can share the same channel if they do not interfere with each other,we ensure that the network remains connected whenever any channel is reclaimed by the authorized user.Particularly,we first introduce successive interference cancellation(SIC)in CRNs to eliminate the interference from the authorized user.Accordingly,,we design a concentric-circle-captured interference protocol to describe the combination of interference cancellation and avoidance.Then,based on the proposed interference protocol,we discover the links between cognitive users that can coexist with the authorized user,and we denote such links as coexisting links by attaching a coexist-weight to them.Next,aiming at minimizing the number of required channels,both centralized and distributed topology control algorithms are proposed to reveal the channelconnected property,and their correctness is verified via theoretical analysis as well.Finally,simulations are conducted to explore the effects of different transmission power levels at the authorized user on the network performance,and show the efficiency of the algorithms on spectrum utilization and robustness.2.We reveal the k-channel-connected property of overlay CRNs.Particularly,a novel definition named k-channel-connectivity is introduced to evaluate the robustness of CRNs.A CRN is referred to as k-channel-connectivity if it could remain connected whenever any k-1(k=2,3,4,...)channels are occupied by authorized users simultaneously.Aiming at solving the k-channel-connectivity guarantee problem,two major issues have to be tackled.On one hand,there exists a tradeoff between guaranteeing k-channel-connectivity and minimizing the number of required channels in a conflict-free CRN.On the other hand,it is also challenging to guarantee k-channel-connectivity with an acceptable computation complexity in a spectrum-scarce CRN.Accordingly,we first design both centralized and distributed topology control algorithms by joint power control and channel assignment.Particularly,we minimize the number of required channels to obtain k-channel-connected property via a successive topology construction strategy.Then,both theoretical analysis and simulation results validate the correctness of proposed algorithms.Meanwhile,simulation results also reveal the performance of our algorithms on constructing different k-channel-connected CRNs.3.We reveal the k-channel-connected property of underlay CRNs.With considering the coexistence of the authorized and cognitive users on the same channel,we ensure the connectivity of the network if any k-1 channels are occupied.Firstly,we adopt SIC for cognitive users to cancel the interference from authorized users;meanwhile,we quantitatively describe the interference or coexistence between authorized and cognitive users.On this basis,aiming at constructing a k-channel-connected CRN,both centralized and distributed algorithms are proposed.Particularly,we design a hybrid topology construction strategy by combining the algorithms proposed above.In addition,under the corresponding network scenario and operating paradigm,the proposed algorithms can be translated to the ones designed in the above two contents,respectively.Finally,the correctness of algorithms is verified via theoretical analysis.Simulation results reveal the effects of the number and transmission power of authorized users on the network performance.