【作者】 赵昆；

【导师】 朱近康；

【作者基本信息】 中国科学技术大学 ， 通信与信息系统， 2006， 博士

【摘要】 论文研究了多输入多输出(MIMO)移动通信系统中的功率控制方法。功率控制的MIMO系统在接收端计算出对信道自适应变化的发送功率，并反馈到发送端，对发送信号作功率加权处理。 论文首先分析了排序串行干扰抵消检测器(OSICD)的子流误比特率(BER)性能差异，提出一种最小化最大的子流BER准则的天线功率控制(PAPC)的方法。该PAPC方法在保持信息速率和总发送功率恒定时，改善OSICD的最差子流的BER性能以提高平均BER性能。论文在收发信号处理中加入对各天线不等功率的考虑，提出了功率控制的OSICD(P-OSICD)。仿真结果说明，最小均方误差准则(MMSE)的P-OSICD可以在小的功率反馈开销下获得与最大似然检测器(MLD)接近的BER性能，并且，P-MMSE-OSICD的计算复杂度相比MMSE-OSICD的增加不大，比MLD的小。 由于OSICD存在数值稳定性不好的问题，论文还研究了一种数值稳定性更好的MIMO检测器排序判决反馈检测器(SDFD)。为了进一步提高SDFD的BER性能，论文提出一种最小化SDFD的BLER性能的PAPC方法，以提高平均BER性能。该PAPC方法使QR分解后的R矩阵具有等对角线元素的性质，同时也最大化自由距离的下限。通过性能分析，论文证明了该PAPC方法是在反馈量和检测性能之间折衷的方法。论文在收发信号处理中加入对各天线不等功率的考虑，提出了功率控制的SDFD(P-SDFD)。仿真结果说明，在接收天线多于发送天线的MIMO系统中，P-SDFD可以在小的功率反馈开销和低的计算复杂度下获得与MLD相似的BER性能。 在蜂窝MIMO移动通信系统中，同信道(CCI)干扰严重恶化了系统性能，论文提出了一种蜂窝MIMO系统中的集中式功率控制(CPC)方法，能够抑制CCI，提高系统的信号干扰比(SIR)中断性能和BER性能。该方法利用了MIMO不同天线发送的数据独立的特点，只调整每个用户发送端的总功率，在用户发送端的不同天线间均分功率的情况下，得到简化的SIR表达式和信道矩阵表达式，并且证明该信道矩阵归一化后满足非负不可约性，可以通过求解归一化的信道矩阵的特征值和特征向量得到平衡的SIR值和相应的发送功率。SIR平衡使系统中最小SIR值最大化。仿真说明了该CPC方法对平均SIR中断概率和平均BER性能的改善。 CPC方法需要收集所有通信链路的信道信息，系统开销大。因此，论文又提出一种蜂窝MIMO系统中的分布式功率控制(DPC)方法，该方法不需要知道系统中所有信道的信息，仅测量目标信号功率和本地SIR值，通过迭代的方法能够收敛于CPC方法得到的发送功率最优解。论文给出了适用于干扰受限系统的DPC算法1和适用于噪声受限系统的DPC算法2，并描述了算法实现框图。DPC算法1的优化目标是最小化平均SIR中断概率；DPC算法2的优化目标是在满足目标SIR的前提下最小化平均发送功率。仿真说明了DPC方法的迭代收敛性和对性能的改善。 功率控制可以提高系统的误比特率(BER)性能，且反馈量不大，计算复杂度的增加也不大。论文所作的研究和贡献，对MIMO移动通信系统逐步从理论走向实用具有重要参考意义。更多还原

【Abstract】 This thesis focuses on the power control approaches in MIMO (Multiple-Input Multiple-Output) mobile communication systems. The power controlled MIMO systems obtain the transmit power which is adaptive to the channel. The transmit power is feedback to the transmitter from the receiver and used to weight the transmit signals.The thesis firstly analyses the substrems’ BER (Bit Error Rate) difference of the OSICD (Ordered Succesive Interference Cancellation Detector) and proposes a PAPC (Per-Antenna Power Control) approach based on the criterion minimizing the substream’s maximum BER. The PAPC approach improves the worst BER performance among substreams to improve the averaged BER performance, while the information rate and total transmit powe keep constant. The thesis takes the unequal antenna power into consideration in MIMO signal process and proposes a P-OSICD (Power controlled OSICD). Simulation results show that the P-MMSE-OSICD (MMSE: Minimum Mean Square Error) can achieve the comparible BER performance to that of the MLD (Maximum Likelihood Detector) at small feedback overhead and lower compelxity. The calculation complexity addition of the P-MMSE-OSICD compared to MMSE-OSICD is small too.Because the numerical stabability of the OSICD is not good, the thesis studies another MIMO detector as SDFD (Sorted Decision Feedback Detector) with better numerical stabability. However, the BER performance of the SDFD is not good enough for real MIMO systems. Hence, the thesis proposes a PAPC approach to improve the BER performance by minimizing the BLER (BLock Error Rate) of SDFD. The PAPC approach makes the matrix R have equal-diagonal elements and maximizes the lower bound of the free distance at the same time. By the performance analysis, the thesis proves that the PAPC approach is a tradeoff between feedback overhead and detection performance. The thesis also proposes a P-SDFD (Power controlled SDFD) concerning the unequl transmit power among transmit antennas. Simulation results show that the P-SDFD achieves the comparible BER performance to that of the MLD at small feedback overhead and lower complexity, when receive antenna number is larger than transmit antenna number.In MIMO cellular mobile communication systems, CCI (Co-Channel Interference) affects the performance greatly. The thesis proposes a CPC (Centralized Power Control) approach to control the CCI and by so to improve the SIR (Signal to Interference Ratio) outage performance and BER performance. The PAPC approach adjusts each user’s total transmit power but keeps queal power levels among antennas of each user. With the fact that signals from multiple transmit antennas are indepedant to each other, a simplified SIR expression and channel matrix expression are derived. The thesis proves that the normalized channel matrix is an irreducible nonnegative matrix. We can obtain the balanced SIR value and the conesponding transmit power by calculation of the eigenvalues and eigenvectors of the normalized channel matrix. SIR balancing maximizes the minimum SIR among users. Simuation results show that the CPC approach can improve the SIR outage probability performance and BER performance.The CPC approach needs to collect channel information of all communication links in the system, which costs much system overhead. The thesis then proposes a DPC (Distributed Power Control) approach to lower the CPC’s overhead. The DPC approach only measures the local desired signal power and local SIR. In iteration fashion, the DPC approach can also achieve the SIR balance. The thesis gives DPC algorithm 1 for interference-limitted systems to minimize the averaged SIR outage更多还原