多用户MIMO-OFDM系统基于波束赋形的接收检测算法研究

发布时间：2019-01-06 18:07

【摘要】：基于子载波共享的多用户MIMO-OFDM系统可以有效改善无线通信系统的传输效率和频谱利用率等性能。但是由于子载波共享方式会产生共道干扰，在一定程度上限制了系统性能的改进。为了在有限的频域资源内提高系统的传输可靠性和频谱利用率等性能，同时降低计算复杂度，本文针对基于子载波共享的多用户MIMO-OFDM系统上行链路的多用户检测问题进行研究。为了提高多用户检测算法性能的同时降低计算复杂度，本文合理结合块对角化技术和串行干扰消除技术，基于矩阵降维理论的零空间交集定理提出具有低复杂度的空域多用户检测算法——基于迭代串行迫零的串行干扰消除算法(ISZF-SIC)。该算法由迭代串行迫零(ISZF)和串行干扰消除(SIC)两部分构成。ISZF基于零空间交集定理，利用块对角化波束赋形技术，通过迭代串行方法，实现了多用户共道干扰抑制；SIC利用ISZF的结果，恰当选择用户检测顺序，通过基于串行干扰消除的空域多用户检测，实现了各用户信息的分离。该算法充分利用了空域资源，相对单独的基于块对角化的检测算法，提高了检测性能和频谱效率等。同时，所提算法通过巧妙的迭代运算，降低了块对角化算法中奇异值分解次数和奇异值分解矩阵维数，极大降低了计算复杂度。提出的ISZF-SIC算法具备检测性能、频谱利用率高和计算复杂度低等优点，但是由于没有考虑用户检测顺序，而是采用随机的检测顺序，，导致当系统信道状况较差时系统性能变差。为了降低非排序产生的差错传播问题，本文在ISZF-SIC算法基础上提出基于信噪比和排序的迭代串行迫零结合串行干扰消除的空域多用户检测算法(order-ISZF-SIC)。该算法的核心是基于SNR排序准则来确定用户的检测顺序，计算用户信噪比的和，选择信噪比和最大的用户作为当前检测的用户。由于根据信道状况确定了合理的用户检测顺序，降低了差错传播概率，从而相对ISZF-SIC算法，order-ISZF-SIC算法进一步改进了系统检测性能和频谱利用率等。同样由于增加了用户排序，相对ISZF-SIC算法，所提排序算法增大了奇异值分解的次数，即增大了计算复杂度。仿真结果表明order-ISZF-SIC算法的性能接近最优穷举排序检测的性能，优于ISZF-SIC算法和基本块对角化算法；其计算复杂度高于ISZF-SIC算法，但是低于基本块对角化算法。
[Abstract]:Multi-user MIMO-OFDM system based on sub-carrier sharing can effectively improve the transmission efficiency and spectrum efficiency of wireless communication systems. However, the performance of the system is limited to a certain extent due to the co-channel interference generated by the sub-carrier sharing mode. In order to improve the transmission reliability and spectrum efficiency of the system in limited frequency domain resources, and reduce the computational complexity, In this paper, the problem of multiuser detection in uplink of multiuser MIMO-OFDM system based on subcarrier sharing is studied. In order to improve the performance of multi-user detection algorithm and reduce the computational complexity, this paper combines block diagonalization with serial interference cancellation. Based on the zero space intersection theorem based on matrix reduction theory, a low complexity spatial multiuser detection algorithm, serial interference cancellation algorithm based on iterative serial zero forcing (ISZF-SIC), is proposed. The algorithm is composed of iterative serial zero forcing (ISZF) and serial interference cancellation (SIC). Based on the zero space intersection theorem, ISZF uses block diagonalization beamforming technology to achieve multiuser co-channel interference suppression by iterative serial method. SIC makes use of the results of ISZF to select the appropriate order of user detection and realizes the separation of user information through spatial multi-user detection based on serial interference cancellation. The algorithm makes full use of spatial resources and improves detection performance and spectral efficiency compared with a single block diagonalization based detection algorithm. At the same time, the proposed algorithm reduces the number of singular value decomposition and the dimension of singular value decomposition matrix through clever iterative operation, and greatly reduces the computational complexity. The proposed ISZF-SIC algorithm has the advantages of high detection performance, high spectral efficiency and low computational complexity. However, due to not considering the order of user detection, the proposed algorithm adopts random detection order. The system performance becomes worse when the channel condition of the system is poor. In order to reduce the problem of error propagation caused by non-sorting, an iterative serial zero-forcing combined with serial interference cancellation based on ISZF-SIC algorithm is proposed in this paper, which is a multi-user detection algorithm in spatial domain (order-ISZF-SIC). The core of the algorithm is to determine the detection order of the user based on the SNR ranking criterion, calculate the sum of the user's SNR, and select the SNR and the maximum user as the users of the current detection. Because the reasonable order of user detection is determined according to the channel condition and the probability of error propagation is reduced, compared with the ISZF-SIC algorithm, the order-ISZF-SIC algorithm further improves the detection performance and spectrum efficiency of the system. The proposed algorithm increases the number of singular value decomposition (SVD) in comparison with the ISZF-SIC algorithm, which increases the computational complexity. Simulation results show that the performance of the order-ISZF-SIC algorithm is close to that of the optimal exhaustive sorting detection and is superior to that of the ISZF-SIC algorithm and the basic block diagonalization algorithm, and its computational complexity is higher than that of the ISZF-SIC algorithm, but lower than that of the basic block diagonalization algorithm.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN929.53

【相似文献】