【作者】 陈力；

【导师】 卫国；

【作者基本信息】 中国科学技术大学 ， 通信与信息系统， 2014， 博士

【摘要】 随着无线通信技术的高速发展,用户对于无线业务的需求也越来越高,这一切都集中反映出了日益增长的无线数据速率要求。这使得现有的无线通信系统受到频谱资源紧缺和能耗严重的双重挑战。随着3G甚至是4G通信网络的广泛架设,可以预见无线通信的能耗将会迅猛增长。作为无线通信系统中能耗的重要组成部分,提高发送端的能效对于降低通信系统的能耗有着重要的现实意义。同时由于移动终端便携性和移动性要求所带来的体积和重量限制,以及电池材料发展的相对缓慢,造成了现有移动通信终端的“电池瓶颈”。因此除了节能环保的意义,提高终端发送能效同样可以延长有效通信时间。因此如何通过终端发送技术,提高终端发送能效是一个亟待解决的问题。本文研究两个终端发送场景：点对点单用户链路和多点对点多用户发送。并分别提出了基于能效的链路自适应发送技术和多用户自适应发送策略,来根据系统的CSIT,干扰状态和目标频谱效率等自适应调整发送速率,发射功率,多天线模式,协作方式等来提高终端发送能效。首先,本文研究在非完美的发送端信道状态信息下,OSTBC-MIMO系统的能效链路自适应问题。之前基于完美信道状态信息的研究和基于统计信道状态信息的研究都可以归为本章研究问题的极限情况。同时,我们不仅仅考虑常值的放大器效率,还考虑MQAM调制时,由于非恒包络调制而造成的与调制阶数相关的放大器效率。尽管建模的问题是非凸的,通过广义凸优化的理论,我们推导出在即时BER约束下,能效最优的发送速率和发射功率的闭式表达结果。根据这个表达结果,我们进一步给出完美信道状态信息和统计信道状态信息下的极限结果。然后我们分别讨论了通过自适应的多天线选择和训练序列优化机制来提高点对点链路的能效。针对能效优化的多天线自适应选择,我们提出了一种基于能效优化的一般性天线选择合并机制,即EE-GSC机制。通过该机制,我们实现了发送端多天线分集增益和电路功耗的最优折中,来最大化链路的能效。基于顺序统计理论,我们推导出了在目标频谱效率下,平均发射天线数和平均发送能效。并基于推导出的理论结果,给出了一些特殊情况的分析。针对能效优化的训练序列自适应功率分配,我们分别分析了基于训练序列的有反馈MIMO系统和无反馈MIMO系统。首先,我们考虑了无反馈MIMO系统中,发送端发送功率在各个天线上平均分配的情况。当训练序列功率固定时,我们给出了能效最优的数据序列功率,同时证明了它的存在唯一性。当数据序列功率和训练序列功率都可变时,我们提出了一种收敛的交替优化算法,获得能效优化的数据序列功率和训练序列功率。接着,我们分析了有反馈的MIMO系统,发送功率在各个天线上注水分配的情况。当训练序列功率固定时,我们给出了最优的数据序列功率和其对应的有效发送天线数。尽管有效发送天线数目是离散的,我们仍证明了它们的存在唯一性。当数据序列功率和训练序列功率都可变时,我们提出了一种类似的收敛的交替优化算法,获得能效优化的数据序列功率和训练序列功率。通过这些算法的数值结果,我们讨论了多天线配置,电路功耗和块衰落长度对于发送端能效的影响。接着我们研究了多用户无用户间协作系统中,用户通过分布式功率控制,在实现目标SINR的约束下,最小化总的发射功耗,以提高用户能效的问题。对于系统中发射用户较少,每个用户都可以达到目标SINR,即系统可行时,我们证明了采用ZF和MMSE接收机,多用户采用标准功控算法时,可以收敛到最优的发射功率。当系统中发射用户较多,或者用户目标SINR较大时,不能保证每个用户都可以达到目标SINR,即系统不可行时,我们提出了具有用户软移除的分布式功控算法。通过广义标准的理论,我们证明了该算法的收敛性。通过数值结果,我们证明了该算法通过软移除不可行用户,不仅仅提高了用户的能效,而且降低了系统中用户的掉线概率。最后我们提出在有用户间协作的MU-SIMO系统中,目标频谱效率下,一种分布式的能效优化体制。我们将该体制分为两个部分。在第一个部分,我们回答这样一个问题,即：“在MU-SIMO协作集合内,用户间如何协作?”。我们给出目标频谱效率下,在MU-SIMO协作集合内,每个用户在各个RB上最优发送能效功率分配的闭式表达式。在第二个部分,我们回答“用户与谁协作,形成MU-SIMO协作集合?”。根据第一部分得到的闭式结果,我们提出了一种基于联盟形成的合作博弈算法,在用户间形成协作的MU-SIMO集合。该算法根据用户能效的Pareto特性,采用了一种融合分裂的收敛迭代操作。更多还原

【Abstract】 With the rapid development of wireless communication technology, users need more and more wireless service. That reflects growing wireless data rate requirements, and the wireless communication system faces challenge of both SE and EE. With the3G and even4G communication networks put into operation, the energy consumption of wireless communication will grow rapidly. As an important part of energy consumption of wireless communication system, reducing the transmit terminal energy consumption has a practical significance. Due to the probability and mobility of users’terminals, the size and weight of them are restricted. Besides the slow progress of battery material, it causes the "battery bottleneck" of users’terminals. Thus, improving the terminals’EE can also improve the effective communication time and prolong the batter life. In a word, how to improve terminals’EE through transmit techniques is a practical problem. In this paper, we concern about energy efficient terminal transmit technique. Two transmit scenes are studied, i.e., point to point link and multi-point to point multi-user transmission. We propose energy efficient link adaption technique and multi-user transmit adaptive strategy, respectively. And the transmitter adaptively changes its transmit rate, transmit power, multi-antennas mode and cooperation mode according to its CSIT, interference state and target SE in order to improve its EE.Firstly, we study energy-efficient link adaptation on Rayleigh Fading channel for OSTBC MIMO system with imperfect CSIT. The former energy-efficient link adaptation works which consider the transmitter with perfect CSI or CDI can be regarded as the limiting cases for this paper. Both modulations with constant PA inefficiency and MQAM with non-constant PA inefficiency are considered in the energy consumption model. The transmission rate and the transmit power are optimized to maximize the EE of the transmitter subject to the I-BER constraint. Although the problem is not concave, we solve it based on the generalized convexity. Closed-form expressions for the most energy-efficient transmission rate and transmit power are given. According to this expression, several special cases such as the transmitter with perfect CSIT and only CDI are discussed, which reveals that the closed-form results are unified.Secondly, we propose an optimal energy efficient GSC (EE-GSC) scheme, which providing a best tradeoff between the diversity gain and circuit power dissipation of multiple antennas, for transmit diversity systems. Based on the classical order statistics results, the average number of active branches with EE-GSC is deduced for the Rayleigh fading scenario. Then the EE performance of EE-GSC scheme is analyzed, and some special cases are also discussed based on the theoretical analysis. We also discuss the EE-optimal power allocation for the training-based MIMO system with and without feedback. Power allocations on training and data signal are discussed to maximize the transmitter’s EE. With pilot power fixed, the EE-optimal data power is deduced, and its existence and uniqueness are also proved. When pilot power and data power are both variables, an iterative and convergent algorithm is proposed to find out suboptimal energy-efficient pilot power and data power.Thirdly, the distributed power control problem is studied in an uplink MU-SIMO scenario. We demonstrate that SPC scheme is also convergent for the uplink MU-SIMO system with the MMSE and ZF receivers which are widely used linear receivers for multi-user detection. When the system becomes infeasible and SPC deteriorates, we propose a gradual soft removal power control (GSR-PC) algorithm which is a unified scheme of SPC and TOPC. It removes the UEs gradually according to the interference they suffer and their tolerance of interference. The GSR-PC algorithm is proved to converge to a unique fix point for the MU-SIMO system with ZF and MMSE receiver. And it reduces the outage ratio through keeping the UEs who can tolerate the interference and removing the users who cannot.Finally, we propose a distributed EE optimization scheme for cooperative MU-SIMO system to achieve each UE’s target SE. We decompose the scheme into two sections. In the first section, we answer the question that "how to cooperate within a MU-SIMO group?" We deduce closed-form expressions of the optimal power and target SE allocations for each UE on each RB within a MU-SIMO group. In the second section, we answer the question that "with whom to form a MU-SIMO group?" According to the deduced expressions, we give a simple algorithm based on the well-known coalition formation game to form MU-SIMO groups among UEs. A convergent iteration of merge-split operations is adopted according to the Pareto order of UEs’EE.更多还原