Researches On Key Techniques Of Beamforming And Detection In LTE Uplink Systems

The development of third-generation mobile communications drives thedevelopment of mobile data business and mobile Internet application, while thedevelopment of the latter poses higher requirements on mobile bandwidth, whichdirectlydrives the development of LTE technology and industry.However, the third-generation mobile communications are facing three problems:(1) Existing virtual MIMO pairing algorithms mainly consider the scenes that theuser terminal equipped with a single antenna. In the communication scene ofthe user terminal equipped with multiple antennas, they cannot make full useof the multi-antenna resources.(2) In order to meet the higher requirements on mobile bandwidth of mobile databusiness and mobile Internet application, existing SINR estimation algorithmsrequire a high complexity in virtual MIMO with more users. It is needed tofind a low complex SINR estimation method.(3) The existing LTE uplink detection method is combination of frequency domainMMSE and Turbo soft interference cancellation, which performance is limited.It is needed to find detection methods with a better performance.This thesis first introduces current research scope and development ofbeamforming algorithms in virtual MIMO systems, SINR estimation algorithms andLTE uplink detection algorithms. This thesis does several exploratory researches on theproblem mentioned above and consists of the following three aspects:First, beamforming algorithm in virtual MIMO systems based on the criterion ofmaximizing SINR is proposed. The basic idea of the algorithm is that for the first layerscheduling user energy is concentrated in the direction of the strongest channel features,the energy of pairing users is concentrated in the direction of the strongest channelcharacteristic and the weakest interference characteristic. For LTE uplink virtual MIMOterminal equipped with multiple antennas, the proposed algorithm makes full use ofantenna resources. Simulation results show that, in LTE uplink under the conditions ofSCME channel, each cell containing20users and two users pairing, virtual MIMO withproposed beamforming algorithm increased the average throughput by11%comparedwith SU-MIMO scheme. Second, low-compexity signal-to-noise ratio estimation and mapping methods areproposed. In order to solve the high-compexity problem of estimating thesignal-to-noise ratio under the MMSE multiuser receiver when multiuser pairing,according to the Hermitian characteristics of the matrix, recursive algorithm is proposedto reduce complexity by using the matrix inversion lemma. Approximation algorithm isproposed to reduce complexity by using the approximate determinat calculation methodof the diagonally dominant matrix. Signal-to-noise ratio mapping method of SC-FDMArequires a large number of division, this thesis presents an approximation algorithmusing inequality, to avoid division operations and reduce complexity.Simulation resultsshow that, in virtual MIMO system with four users pairing, as compared to the existingalgorithm, the proposed recursive algorithm and approximation algorithm reducescomplexity by59.97%and73.79%when the base station equipped four antennas, andby50.65%and61.99%when the base station equipped eight antennas.Third, MSSFE algorithm is proposed to improve the existing SSFE algorithm.Against the LTE PUSCH, two kinds of time-frequency dual-domain equalizationreceiver are proposed. The basic idea is that to suppress the weak interferencesuppression in frequency domain by MMSE equalizer, and in the time domain to jointdetect the strong interference between users of virtual MIMO. Simulation results showthat, in a4×4MIMO-OFDM system,16QAM modulation parameter m=[1,2,4,8], theproposed MSSFE algorithm outperform the original SSFE algorithm by3dB. Extendedpedestrian A model is used in simulations, and each user possess five resource blocks.The proposed MMSE+SD and MMSE+SDE schemes without iterations outperformMMSE equalizer by2.5dB and1dB. After the three turbo PIC iterations, the proposedMMSE+SD and MMSE+SDE schemes obtains2dB and1dB gain over theconventional algorithm.In this thesis, beamforming algorithms in virtual MIMO systems, SINR estimationalgorithms and LTE uplink detection algorithms have been analysed. The proposedvirtual MIMO beamforming algorithm based on maximum SINR criteria, which canmake full use of resources of multiple antennas, can be applied when the terminal isequipped with mutiple antennas. The SINR estimation methods proposed in this thesis,which can reduce the computational complexity, can be applied in a virtual MIMOsystem. The time-frequency dual-domain receiver schemes for SC-FDMA system areproposed to improve the detection performance and can be applied to LTE, LTE-Asystem and other communication system using the SC-FDMA technology.