| With the rapid development of wireless communications,a growing demand for wireless radio spectrum is needed,which leads to the scarcity of the available spectrum resource.Thus,the so-called“spectrum crisis”begins to spread all over the world.On the other hand,the spectrum policy task force of Federal Communications Commission reports that a large portion of the assigned spectrum remains underutilized.Therefore,Cognitive Radio(CR)technology which adopts the dynamic spectrum access scheme,was proposed as a promising solution.The basic idea of CR is spectrum sharing or spectrum reusing.To achieve this,the cognitive users(CUs)are required to operate the spectrum sensing frequently,that is,the detection of spec-trum holes which are not occupied by the primary users(PUs).Hence,spectrum sensing is the fundamental and main task of CR.Conventional spectrum sensing methods mainly focus on the time,frequency and geography dimensions,which have limit improvement of spectrum efficien-cy.Meanwhile,the rapid development of multi-antenna technology makes the terminal obtain the angle recognition capability,which contributes to the studies of spectrum sensing in spatial dimension.In this case,the CUs can make spectrum access by avoiding the directions of PUs’communications,which brings additional spatial multiplexing gain for spectrum opportunity and thus further improves the spectrum efficiency.The work of this dissertation focuses on the spatial spectrum sensing and makes a thorough analysis.Utilizing the feature that the spatial spectrum carries the spatial information itself,a novel spatial spectrum sensing framework is proposed.Under the proposed framework,we study the sensing schemes,sensing algorithms and the prac-tical sensing-throughput tradeoff problem.The main contributions of this dissertation are shown as followings.Firstly,we use the Angle of Arrival(AoA)estimation technology to obtain the continues-form spatial spectrum,and design the AoA estimation based spatial spectrum sensing algorithms.By using the AoA estimation technology to make spatial filtering,we can analyze the problem of spectrum sensing from the aspect of spatial spectrum in angle domain.Taking advantage of the differences of spatial spectrum distribution,we design the AoA estimation based spatial spectrum sensing scheme.Under the proposed scheme,we first propose the spectrum-value ratio based detection algorithms using feature of PUs’ spatial spectrum;and we further study the coherent feature of noise’s spatial spectrum to design a central symmetry based feature detection algorithm.In addition,the corresponding theoretical analysis is also provided.Simulation results show that the proposed methods can achieve higher probability of detection by capturing the features of spatial spectrum,and they can also provide the available AoA information.Secondly,we focus on the discrete-form spatial spectrum which is attainted through beam-forming technology,and propose beamforming based spatial spectrum sensing algorithms.Uti-lizing beamforming technology to make space divided into sectors,we can then transfer the is-sue of spectrum sensing into the detection of discrete-form sectorized spatial spectrum.Through analyzing the energy differences among the sectors,we design the beamforming based spatial spectrum sensing scheme.Under the proposed scheme,we first exploit the feature of energy distribution to propose energy ratio based detection methods;and then we use the theory of like-lihood ratio test to develop the energy weighting based detection algorithms.In addition,the corresponding theoretical analysis is also provided.The proposed algorithms are able to capture the spatial diversity gain to obtain better detection performance;meanwhile they can provide the available sectors for spectrum access.Simulation results are presented to verify the efficiency of the proposed algorithms.Finally,the work turns to the performance analysis of spatial spectrum sensing and stud-ies the spatial spectrum based sensing-throughput tradeoff scheme.In this work,the spatial-temporal uncertainty principle(STUP)is analyzed and derived,which reveals an interesting constraint phenomenon between spatial and temporal resolutions.The proposed STUP is fur-ther studied and then the spatial-temporal resolution tradeoff(STRT)scheme is proposed,in which the sensing-throughput tradeoff problem is modelled as a kind of“spatial multiplexing gain-access time”tradeoff problem.The proposed STRT scheme analyzes and proves that there exists an optimal spatial resolution.In addition,a heuristic search algorithm is proposed to track the optimal spatial resolution at an exponential convergence rate.Simulation results are present-ed to verify the feasibility of the proposed schemes.To sum up,the above works make a valuable exploration on the sensing algorithms and per-formance analysis for CR networks,which enriches the related theory of spatial spectrum sensing methods,improves the feasibility in real systems and has a certain theoretical significance and practical value. |
