Research On Key Techniques Of Cooperative Spectrum Sensing In Cognitive Radio Systems

The rapid development of radio technology gives rise to an explosively increasing demand for the spectrum resources,which makes the problem of spectrum scarcity increasingly serious,thus severely restricting the development of wireless communication technology.On the other hand,under the current static spectrum resource management and allocation strategy,the allocated authorized spectrum exhibit varying degrees of idleness in multiple dimensions,and the spectrum efficiency is low.In this case,cognitive radio technology arises at the historic moment.As long as it does not interfere with authorized users,cognitive users are allowed to access the idle resources in authorized spectrum by the means of borrowing,so as to achieve the purpose of alleviating the lack of spectrum resources.As the most basic and the most critical part of cognitive radio,spectrum sensing technology has attracted extensive attention in the academic world.In order to ensure that authorized users are not interfered by cognitive users and provide more opportunities to cognitive users,spectrum sensing technology is required to have better accuracy and higher efficiency.To account for this,this dissertation carries out in-depth research and analysis on the spectrum sensing problem in cognitive radio system.The main work and contributions are as follows:At first,for local and cooperative sensing,we evaluate the sensing performances over Time-Correlated Rayleigh Channels in Mobile Environments.The closed expression of detection probability for both static Rayleigh channel and fully independent Rayleigh channel,which serve as a upper and lower bounds for the detection probability for spectrum sensing over correlated Rayleigh fading channel,and the simulation results validate the validity of the developed theoretical analysis.Secondly,we focus on the research of cooperative spectrum sensing based on hard fusion in mobile environment.Considering that the sensing channels of different cognitive users have different large-scale fading,we propose a hard fusion algorithm for ideal reporting channels.The algorithm weights the hard decisions based on the user's large-scale fading gain ordering,and has better detection performance than the traditional equal-gain algorithm,and does not need to know the accurate large-scale fading gain.Further,we study collaborative spectrum sensing algorithm based on hard fusion in mobile non-ideal reporting channels,and investigate the effect of time correlation on collaborative sensing based on hard fusion.At the same time,the influence of bit error rate on collaborative sensing performance is analyzed when the local sensing information of cognitive users passes through the small scale fading reporting channel.The approximate closed expressions of false alarm and detection probability of the system are derived,and the theoretical analysis results are verified by simulation results.Thirdly,in view of the contradiction between collaborative sensing performance and reporting overhead,a collaborative spectrum sensing algorithm based on multi-bit quantization with known average signal noise radio is proposed.Simulation results show that compared with collaborative sensing based on hard fusion,the proposed algorithm can greatly improve the spectrum sensing performance on the premise that each sensing node only increases the reporting overhead of one bit;Compared with collaborative sensing based on soft fusion,this algorithm has similar detection performance as soft fusion,but requires much less reporting overhead.On this basis,an improved throughput-efficient collaborative sensing algorithm is proposed.The setting of local sensing time and reporting time in the collaborative sensing process directly affects the accuracy of the global decision result and is related to the throughput of cognitive system.The improved algorithm reduces the required reporting overhead by using cooperative sensing with two bits quantization,and extends the spectrum sampling operation from the traditional sense of local sensing time to other cognitive users' reporting time.In order to further improve the detection performance and effective throughput,considering the difference of the channels of each cognitive user in the cognitive network,the user with better sensing condition is arranged at the back so that it has a longer sampling window.The simulation results show that the proposed scheme is superior to the traditional sensing scheme in both sensing accuracy and effective throughput,and the proposed scheme has a lower complexity.Finally,in order to solve the problem of excessive energy consumption,an energyefficient cooperative spectrum sensing algorithm in cluster-based cognitive networks is proposed.By clustering,the report energy consumption of a single user is reduced.In order to further reduce the sensing energy consumption,a two-step collaborative spectrum sensing algorithm is proposed.In the first step,only the cluster head performs single-node spectrum sensing,and the second step is triggered when the cluster head cannot determine the existence of the authorized user.The cluster head enables the cognitive users in the cluster to participate in the collaboration and jointly sensing the existence of the authorized user.The intra-cluster collaborative spectrum sensing is triggered only when necessary,and the quantitative data fusion scheme with less than two bits is adopted after the trigger.Simulation results demonstrate the effectiveness of the proposed algorithm from three aspects:detection accuracy,throughput and energy efficiency.