Detection For Analog Multiband Signals On A Wide Band Based On Compressive Sensing

International specification of ITU IMT-Advanced,which is the 4th generation wirelss communications system, is started in 2008. The work is much attendtioned on a world wide scale.The future IMT-Advanced system will support wireless transmission of multiband signals on a wide band.The system can reach a peak rate of 1Giga bits per second and cover a spectral range from 450MHz to 3.6GHz in which there are several discrete widebands(20MHz from 450 MHz to 470 MHz,108MHz from 698MHz to 806MHz,100MHz from 2.3GHz to 2.4 GHz,200MHz from 3.4GHz to 3.6 GHz).When wide multiband signals are detected by the traditional method which is based on the present Nyquist sampling theorem,the difficulty of RF unit design, the complexity of RF unit,the requirement of hardware and the costs will all be greatly increased. Therefore, we urgently need a new theory or technology to solve the problem of wide multiband signals detection. So how to effectively realize the detection for multiband signals on a wide band is a important problem which the implementation of physical layer in 4th generation wireless communications system is faced with. The "Compressive Sensing" theory is a new theory proposed in recent years. It can breakthrough restrictions of SyQuest sampling theorem when capturing data and pave a new way for data storage, transmission and processing. So it will be the most potential theory to do wide multiband signals detection. The thesis firstly explains the background of newly born compressive sensing theory and describes the problem it aims to solve. Then it summarizes the present compressive sensing technology in both discrete and analog domain. Detailed explanation is presented on the signals'sparse representation, measurement matrix design and reconstruction algorithm design in the part of discrete signals' compressive sensing. In the part of analog signals'compressive sensing, it gives discussion on mainstream implementation method of MWC and AIC and analyses their sampling method and recovery principles.On the basis of theoretical foundation above, a sub-sampling detection structure and jointly recovery algorithms based on multi-antenna compressive sensing for wide multiband signals are proposed in this thesis, which contains CRL2 algorithm and CBS algorithm。The structure can realize wide multiband signal detection and the proposed jointly recovery algorithms based on muti-antenna can improve the detection performance compared with single antenna structure. What's more, based on the proposed structure and algorithms, this thesis gives two application scenarios of compressive sensing which are wide mutiband spectrum sensing and wireless localization. Besides, the simulations are presented to verify their feasibility. Since wide multiband signal detection is a issue in future 4G wireless communication system which is represented by Gbps wireless communication system, the thesis finally introduces the TDD-MIMO-OFDM wireless communication key technology verification demo system developed by author's R&D group, which can reach peak transmisstion rate of 1Gbps. This part focuses on the parameters design according to system targets and implementation of MIMO detection.